qemu_driver.c 197.1 KB
Newer Older
D
Daniel P. Berrange 已提交
1 2 3
/*
 * driver.c: core driver methods for managing qemu guests
 *
4
 * Copyright (C) 2006, 2007, 2008, 2009 Red Hat, Inc.
D
Daniel P. Berrange 已提交
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
 * Copyright (C) 2006 Daniel P. Berrange
 *
 * This library is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 * This library is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with this library; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307  USA
 *
 * Author: Daniel P. Berrange <berrange@redhat.com>
 */

24
#include <config.h>
25

D
Daniel P. Berrange 已提交
26 27 28 29 30 31 32 33 34 35 36
#include <sys/types.h>
#include <sys/poll.h>
#include <dirent.h>
#include <limits.h>
#include <string.h>
#include <stdio.h>
#include <strings.h>
#include <stdarg.h>
#include <stdlib.h>
#include <unistd.h>
#include <errno.h>
37
#include <sys/utsname.h>
38 39 40 41
#include <sys/stat.h>
#include <fcntl.h>
#include <signal.h>
#include <paths.h>
42 43
#include <pwd.h>
#include <stdio.h>
44
#include <sys/wait.h>
45
#include <sys/ioctl.h>
D
Daniel P. Berrange 已提交
46

47 48 49 50
#if HAVE_SCHED_H
#include <sched.h>
#endif

51
#include "virterror_internal.h"
52
#include "logging.h"
53
#include "datatypes.h"
54 55
#include "qemu_driver.h"
#include "qemu_conf.h"
J
Jim Meyering 已提交
56
#include "c-ctype.h"
57
#include "event.h"
58
#include "buf.h"
59
#include "util.h"
60
#include "nodeinfo.h"
61
#include "stats_linux.h"
62
#include "capabilities.h"
63
#include "memory.h"
64
#include "uuid.h"
65
#include "domain_conf.h"
66 67
#include "node_device_conf.h"
#include "pci.h"
68
#include "security.h"
69
#include "cgroup.h"
70

71

72 73
#define VIR_FROM_THIS VIR_FROM_QEMU

R
Richard W.M. Jones 已提交
74
/* For storing short-lived temporary files. */
75
#define TEMPDIR LOCAL_STATE_DIR "/cache/libvirt/qemu"
R
Richard W.M. Jones 已提交
76

77 78 79
#define QEMU_CMD_PROMPT "\n(qemu) "
#define QEMU_PASSWD_PROMPT "Password: "

80 81
static int qemudShutdown(void);

82 83
static void qemuDriverLock(struct qemud_driver *driver)
{
84
    virMutexLock(&driver->lock);
85 86 87
}
static void qemuDriverUnlock(struct qemud_driver *driver)
{
88
    virMutexUnlock(&driver->lock);
89 90
}

91 92 93
static void qemuDomainEventFlush(int timer, void *opaque);
static void qemuDomainEventQueue(struct qemud_driver *driver,
                                 virDomainEventPtr event);
94

95 96
static void qemudDispatchVMEvent(int watch,
                                 int fd,
97 98 99
                                 int events,
                                 void *opaque);

100 101
static int qemudStartVMDaemon(virConnectPtr conn,
                              struct qemud_driver *driver,
102
                              virDomainObjPtr vm,
103 104
                              const char *migrateFrom,
                              int stdin_fd);
105

106 107
static void qemudShutdownVMDaemon(virConnectPtr conn,
                                  struct qemud_driver *driver,
108
                                  virDomainObjPtr vm);
109

110
static int qemudDomainGetMaxVcpus(virDomainPtr dom);
111

112 113 114 115 116 117 118 119
static int qemudMonitorCommand(const virDomainObjPtr vm,
                               const char *cmd,
                               char **reply);
static int qemudMonitorCommandExtra(const virDomainObjPtr vm,
                                    const char *cmd,
                                    const char *extra,
                                    const char *extraPrompt,
                                    char **reply);
120 121 122
static int qemudDomainSetMemoryBalloon(virConnectPtr conn,
                                       virDomainObjPtr vm,
                                       unsigned long newmem);
123 124
static int qemudDetectVcpuPIDs(virConnectPtr conn,
                               virDomainObjPtr vm);
125

J
Jim Meyering 已提交
126
static struct qemud_driver *qemu_driver = NULL;
127 128


129
static int
130
qemudLogFD(virConnectPtr conn, struct qemud_driver *driver, const char* name)
131 132 133
{
    char logfile[PATH_MAX];
    mode_t logmode;
G
Guido Günther 已提交
134
    int ret, fd = -1;
135

136 137
    if ((ret = snprintf(logfile, sizeof(logfile), "%s/%s.log",
                        driver->logDir, name))
G
Guido Günther 已提交
138
        < 0 || ret >= sizeof(logfile)) {
139
        virReportOOMError(conn);
140 141 142 143
        return -1;
    }

    logmode = O_CREAT | O_WRONLY;
144 145
    /* Only logrotate files in /var/log, so only append if running privileged */
    if (driver->privileged)
146
        logmode |= O_APPEND;
147 148 149
    else
        logmode |= O_TRUNC;

150
    if ((fd = open(logfile, logmode, S_IRUSR | S_IWUSR)) < 0) {
151 152 153
        virReportSystemError(conn, errno,
                             _("failed to create logfile %s"),
                             logfile);
154 155
        return -1;
    }
156
    if (virSetCloseExec(fd) < 0) {
157 158
        virReportSystemError(conn, errno, "%s",
                             _("Unable to set VM logfile close-on-exec flag"));
159 160 161 162 163 164 165
        close(fd);
        return -1;
    }
    return fd;
}


166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182
static int
qemudLogReadFD(virConnectPtr conn, const char* logDir, const char* name, off_t pos)
{
    char logfile[PATH_MAX];
    mode_t logmode = O_RDONLY;
    int ret, fd = -1;

    if ((ret = snprintf(logfile, sizeof(logfile), "%s/%s.log", logDir, name))
        < 0 || ret >= sizeof(logfile)) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("failed to build logfile name %s/%s.log"),
                         logDir, name);
        return -1;
    }


    if ((fd = open(logfile, logmode)) < 0) {
183 184 185
        virReportSystemError(conn, errno,
                             _("failed to create logfile %s"),
                             logfile);
186 187
        return -1;
    }
188
    if (virSetCloseExec(fd) < 0) {
189 190
        virReportSystemError(conn, errno, "%s",
                             _("Unable to set VM logfile close-on-exec flag"));
191 192 193 194
        close(fd);
        return -1;
    }
    if (lseek(fd, pos, SEEK_SET) < 0) {
195 196 197
        virReportSystemError(conn, errno,
                             _("Unable to seek to %lld in %s"),
                             (long long) pos, logfile);
198 199 200 201 202 203
        close(fd);
    }
    return fd;
}


204 205 206
static void
qemudAutostartConfigs(struct qemud_driver *driver) {
    unsigned int i;
207 208 209 210 211
    /* XXX: Figure out a better way todo this. The domain
     * startup code needs a connection handle in order
     * to lookup the bridge associated with a virtual
     * network
     */
212 213 214
    virConnectPtr conn = virConnectOpen(driver->privileged ?
                                        "qemu:///system" :
                                        "qemu:///session");
215
    /* Ignoring NULL conn which is mostly harmless here */
216

217
    qemuDriverLock(driver);
218
    for (i = 0 ; i < driver->domains.count ; i++) {
219
        virDomainObjPtr vm = driver->domains.objs[i];
220
        virDomainObjLock(vm);
221 222
        if (vm->autostart &&
            !virDomainIsActive(vm)) {
223
            int ret = qemudStartVMDaemon(conn, driver, vm, NULL, -1);
224 225
            if (ret < 0) {
                virErrorPtr err = virGetLastError();
226 227
                VIR_ERROR(_("Failed to autostart VM '%s': %s\n"),
                          vm->def->name,
228
                          err ? err->message : "");
229
            } else {
230 231 232 233 234 235
                virDomainEventPtr event =
                    virDomainEventNewFromObj(vm,
                                             VIR_DOMAIN_EVENT_STARTED,
                                             VIR_DOMAIN_EVENT_STARTED_BOOTED);
                if (event)
                    qemuDomainEventQueue(driver, event);
236
            }
237
        }
238
        virDomainObjUnlock(vm);
239
    }
240
    qemuDriverUnlock(driver);
241

242 243
    if (conn)
        virConnectClose(conn);
244 245
}

246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262

/**
 * qemudRemoveDomainStatus
 *
 * remove all state files of a domain from statedir
 *
 * Returns 0 on success
 */
static int
qemudRemoveDomainStatus(virConnectPtr conn,
                        struct qemud_driver *driver,
                        virDomainObjPtr vm)
{
    int rc = -1;
    char *file = NULL;

    if (virAsprintf(&file, "%s/%s.xml", driver->stateDir, vm->def->name) < 0) {
263
        virReportOOMError(conn);
264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282
        goto cleanup;
    }

    if (unlink(file) < 0 && errno != ENOENT && errno != ENOTDIR) {
        qemudReportError(conn, vm, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("Failed to unlink status file %s"), file);
        goto cleanup;
    }

    if(virFileDeletePid(driver->stateDir, vm->def->name))
        goto cleanup;

    rc = 0;
cleanup:
    VIR_FREE(file);
    return rc;
}


283 284 285 286 287
static int qemudOpenMonitor(virConnectPtr conn,
                            struct qemud_driver* driver,
                            virDomainObjPtr vm,
                            int reconnect);

288 289 290 291

/*
 * Open an existing VM's monitor, re-detect VCPU threads
 * and re-reserve the security labels in use
292 293
 */
static int
294 295
qemuReconnectDomain(struct qemud_driver *driver,
                    virDomainObjPtr obj)
296
{
297
    int rc;
298

299
    if ((rc = qemudOpenMonitor(NULL, driver, obj, 1)) != 0) {
300 301 302 303
        VIR_ERROR(_("Failed to reconnect monitor for %s: %d\n"),
                  obj->def->name, rc);
        goto error;
    }
304

305 306 307
    if (qemudDetectVcpuPIDs(NULL, obj) < 0) {
        goto error;
    }
308

309 310 311 312 313
    if (obj->def->seclabel.type == VIR_DOMAIN_SECLABEL_DYNAMIC &&
        driver->securityDriver &&
        driver->securityDriver->domainReserveSecurityLabel &&
        driver->securityDriver->domainReserveSecurityLabel(NULL, obj) < 0)
        return -1;
314

315 316
    if (obj->def->id >= driver->nextvmid)
        driver->nextvmid = obj->def->id + 1;
317

318
    return 0;
319

320 321 322
error:
    return -1;
}
323

324 325 326 327 328 329 330 331 332 333
/**
 * qemudReconnectVMs
 *
 * Try to re-open the resources for live VMs that we care
 * about.
 */
static void
qemuReconnectDomains(struct qemud_driver *driver)
{
    int i;
334

335 336 337 338 339 340 341 342 343
    for (i = 0 ; i < driver->domains.count ; i++) {
        virDomainObjPtr obj = driver->domains.objs[i];

        virDomainObjLock(obj);
        if (qemuReconnectDomain(driver, obj) < 0) {
            /* If we can't get the monitor back, then kill the VM
             * so user has ability to start it again later without
             * danger of ending up running twice */
            qemudShutdownVMDaemon(NULL, driver, obj);
344
        }
345
        virDomainObjUnlock(obj);
346 347 348
    }
}

349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381

static int
qemudSecurityCapsInit(virSecurityDriverPtr secdrv,
                      virCapsPtr caps)
{
    const char *doi, *model;

    doi = virSecurityDriverGetDOI(secdrv);
    model = virSecurityDriverGetModel(secdrv);

    caps->host.secModel.model = strdup(model);
    if (!caps->host.secModel.model) {
        char ebuf[1024];
        VIR_ERROR(_("Failed to copy secModel model: %s"),
                  virStrerror(errno, ebuf, sizeof ebuf));
        return -1;
    }

    caps->host.secModel.doi = strdup(doi);
    if (!caps->host.secModel.doi) {
        char ebuf[1024];
        VIR_ERROR(_("Failed to copy secModel DOI: %s"),
                  virStrerror(errno, ebuf, sizeof ebuf));
        return -1;
    }

    VIR_DEBUG("Initialized caps for security driver \"%s\" with "
              "DOI \"%s\"", model, doi);

    return 0;
}


382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401
static int
qemudSecurityInit(struct qemud_driver *qemud_drv)
{
    int ret;
    virSecurityDriverPtr security_drv;

    ret = virSecurityDriverStartup(&security_drv,
                                   qemud_drv->securityDriverName);
    if (ret == -1) {
        VIR_ERROR0(_("Failed to start security driver"));
        return -1;
    }
    /* No security driver wanted to be enabled: just return */
    if (ret == -2) {
        VIR_INFO0(_("No security driver available"));
        return 0;
    }

    qemud_drv->securityDriver = security_drv;

402
    VIR_INFO("Initialized security driver %s", security_drv->name);
403 404 405 406 407

    /*
     * Add security policy host caps now that the security driver is
     * initialized.
     */
408 409
    return qemudSecurityCapsInit(security_drv, qemud_drv->caps);
}
410 411


412

413 414 415 416 417 418
/**
 * qemudStartup:
 *
 * Initialization function for the QEmu daemon
 */
static int
419
qemudStartup(int privileged) {
420
    char *base = NULL;
D
Daniel P. Berrange 已提交
421
    char driverConf[PATH_MAX];
422
    int rc;
423

424
    if (VIR_ALLOC(qemu_driver) < 0)
425 426
        return -1;

427
    if (virMutexInit(&qemu_driver->lock) < 0) {
428
        VIR_ERROR("%s", _("cannot initialize mutex"));
429 430 431
        VIR_FREE(qemu_driver);
        return -1;
    }
432
    qemuDriverLock(qemu_driver);
433
    qemu_driver->privileged = privileged;
434

435 436 437
    /* Don't have a dom0 so start from 1 */
    qemu_driver->nextvmid = 1;

438 439
    /* Init callback list */
    if(VIR_ALLOC(qemu_driver->domainEventCallbacks) < 0)
440
        goto out_of_memory;
441 442 443 444 445 446
    if (!(qemu_driver->domainEventQueue = virDomainEventQueueNew()))
        goto out_of_memory;

    if ((qemu_driver->domainEventTimer =
         virEventAddTimeout(-1, qemuDomainEventFlush, qemu_driver, NULL)) < 0)
        goto error;
447

448
    if (privileged) {
449 450
        if (virAsprintf(&qemu_driver->logDir,
                        "%s/log/libvirt/qemu", LOCAL_STATE_DIR) == -1)
451
            goto out_of_memory;
452

D
Daniel P. Berrange 已提交
453
        if ((base = strdup (SYSCONF_DIR "/libvirt")) == NULL)
454
            goto out_of_memory;
455 456 457 458

        if (virAsprintf(&qemu_driver->stateDir,
                      "%s/run/libvirt/qemu/", LOCAL_STATE_DIR) == -1)
            goto out_of_memory;
459
    } else {
460
        uid_t uid = geteuid();
461 462
        char *userdir = virGetUserDirectory(NULL, uid);
        if (!userdir)
463
            goto error;
464

465
        if (virAsprintf(&qemu_driver->logDir,
466 467
                        "%s/.libvirt/qemu/log", userdir) == -1) {
            VIR_FREE(userdir);
468
            goto out_of_memory;
469
        }
470

471 472
        if (virAsprintf(&base, "%s/.libvirt", userdir) == -1) {
            VIR_FREE(userdir);
473
            goto out_of_memory;
474 475
        }
        VIR_FREE(userdir);
476 477 478 479 480 481

        if (virAsprintf(&qemu_driver->stateDir, "%s/qemu/run", base) == -1)
            goto out_of_memory;
    }

    if (virFileMakePath(qemu_driver->stateDir) < 0) {
482
        char ebuf[1024];
483 484
        VIR_ERROR(_("Failed to create state dir '%s': %s\n"),
                  qemu_driver->stateDir, virStrerror(errno, ebuf, sizeof ebuf));
485
        goto error;
486 487 488 489 490
    }

    /* Configuration paths are either ~/.libvirt/qemu/... (session) or
     * /etc/libvirt/qemu/... (system).
     */
D
Daniel P. Berrange 已提交
491
    if (snprintf (driverConf, sizeof(driverConf), "%s/qemu.conf", base) == -1)
492
        goto out_of_memory;
D
Daniel P. Berrange 已提交
493
    driverConf[sizeof(driverConf)-1] = '\0';
494

495
    if (virAsprintf(&qemu_driver->configDir, "%s/qemu", base) == -1)
496 497
        goto out_of_memory;

498
    if (virAsprintf(&qemu_driver->autostartDir, "%s/qemu/autostart", base) == -1)
499 500
        goto out_of_memory;

501
    VIR_FREE(base);
502

503 504 505 506 507 508 509
    rc = virCgroupForDriver("qemu", &qemu_driver->cgroup, privileged, 1);
    if (rc < 0) {
        char buf[1024];
        VIR_WARN("Unable to create cgroup for driver: %s",
                 virStrerror(-rc, buf, sizeof(buf)));
    }

510 511
    if ((qemu_driver->caps = qemudCapsInit()) == NULL)
        goto out_of_memory;
D
Daniel P. Berrange 已提交
512

513
    if (qemudLoadDriverConfig(qemu_driver, driverConf) < 0) {
514 515 516
        goto error;
    }

517
    if (qemudSecurityInit(qemu_driver) < 0) {
518
        goto error;
D
Daniel P. Berrange 已提交
519 520
    }

521 522 523 524 525 526 527 528 529 530 531 532
    /* Get all the running persistent or transient configs first */
    if (virDomainLoadAllConfigs(NULL,
                                qemu_driver->caps,
                                &qemu_driver->domains,
                                qemu_driver->stateDir,
                                NULL,
                                1, NULL, NULL) < 0)
        goto error;

    qemuReconnectDomains(qemu_driver);

    /* Then inactive persistent configs */
533 534 535 536
    if (virDomainLoadAllConfigs(NULL,
                                qemu_driver->caps,
                                &qemu_driver->domains,
                                qemu_driver->configDir,
537
                                qemu_driver->autostartDir,
538
                                0, NULL, NULL) < 0)
539
        goto error;
540 541
    qemuDriverUnlock(qemu_driver);

542 543
    qemudAutostartConfigs(qemu_driver);

544

545 546
    return 0;

547
out_of_memory:
548
    virReportOOMError(NULL);
549 550 551
error:
    if (qemu_driver)
        qemuDriverUnlock(qemu_driver);
552
    VIR_FREE(base);
553
    qemudShutdown();
554 555 556
    return -1;
}

557 558 559 560
static void qemudNotifyLoadDomain(virDomainObjPtr vm, int newVM, void *opaque)
{
    struct qemud_driver *driver = opaque;

561 562 563 564 565 566 567 568
    if (newVM) {
        virDomainEventPtr event =
            virDomainEventNewFromObj(vm,
                                     VIR_DOMAIN_EVENT_DEFINED,
                                     VIR_DOMAIN_EVENT_DEFINED_ADDED);
        if (event)
            qemuDomainEventQueue(driver, event);
    }
569 570
}

571 572 573 574 575 576 577 578
/**
 * qemudReload:
 *
 * Function to restart the QEmu daemon, it will recheck the configuration
 * files and update its state and the networking
 */
static int
qemudReload(void) {
579 580 581
    if (!qemu_driver)
        return 0;

582
    qemuDriverLock(qemu_driver);
583 584 585 586
    virDomainLoadAllConfigs(NULL,
                            qemu_driver->caps,
                            &qemu_driver->domains,
                            qemu_driver->configDir,
587
                            qemu_driver->autostartDir,
588
                            0, qemudNotifyLoadDomain, qemu_driver);
589
    qemuDriverUnlock(qemu_driver);
590

591
    qemudAutostartConfigs(qemu_driver);
592 593

    return 0;
594 595
}

596 597 598 599 600 601 602 603 604 605
/**
 * qemudActive:
 *
 * Checks if the QEmu daemon is active, i.e. has an active domain or
 * an active network
 *
 * Returns 1 if active, 0 otherwise
 */
static int
qemudActive(void) {
606
    unsigned int i;
607
    int active = 0;
608

609 610 611
    if (!qemu_driver)
        return 0;

612 613 614 615 616 617 618 619
    qemuDriverLock(qemu_driver);
    for (i = 0 ; i < qemu_driver->domains.count ; i++) {
        virDomainObjPtr vm = qemu_driver->domains.objs[i];
        virDomainObjLock(vm);
        if (virDomainIsActive(vm))
            active = 1;
        virDomainObjUnlock(vm);
    }
620

621 622
    qemuDriverUnlock(qemu_driver);
    return active;
623 624
}

625 626 627 628 629 630 631
/**
 * qemudShutdown:
 *
 * Shutdown the QEmu daemon, it will stop all active domains and networks
 */
static int
qemudShutdown(void) {
632

633
    if (!qemu_driver)
634
        return -1;
635

636
    qemuDriverLock(qemu_driver);
637 638
    virCapabilitiesFree(qemu_driver->caps);

639
    virDomainObjListFree(&qemu_driver->domains);
640

641
    VIR_FREE(qemu_driver->securityDriverName);
642
    VIR_FREE(qemu_driver->logDir);
643 644
    VIR_FREE(qemu_driver->configDir);
    VIR_FREE(qemu_driver->autostartDir);
645
    VIR_FREE(qemu_driver->stateDir);
646
    VIR_FREE(qemu_driver->vncTLSx509certdir);
J
Jim Meyering 已提交
647
    VIR_FREE(qemu_driver->vncListen);
648
    VIR_FREE(qemu_driver->vncPassword);
649
    VIR_FREE(qemu_driver->vncSASLdir);
D
Daniel P. Berrange 已提交
650

651 652
    /* Free domain callback list */
    virDomainEventCallbackListFree(qemu_driver->domainEventCallbacks);
653 654 655 656
    virDomainEventQueueFree(qemu_driver->domainEventQueue);

    if (qemu_driver->domainEventTimer != -1)
        virEventRemoveTimeout(qemu_driver->domainEventTimer);
657

658 659 660
    if (qemu_driver->brctl)
        brShutdown(qemu_driver->brctl);

661 662
    virCgroupFree(&qemu_driver->cgroup);

663
    qemuDriverUnlock(qemu_driver);
664
    virMutexDestroy(&qemu_driver->lock);
665
    VIR_FREE(qemu_driver);
666 667

    return 0;
668 669 670
}

/* Return -1 for error, 1 to continue reading and 0 for success */
671
typedef int qemudHandlerMonitorOutput(virConnectPtr conn,
672
                                      virDomainObjPtr vm,
673 674 675
                                      const char *output,
                                      int fd);

676 677 678
/*
 * Returns -1 for error, 0 on end-of-file, 1 for success
 */
679
static int
680
qemudReadMonitorOutput(virConnectPtr conn,
681
                       virDomainObjPtr vm,
682 683
                       int fd,
                       char *buf,
G
Guido Günther 已提交
684
                       size_t buflen,
685
                       qemudHandlerMonitorOutput func,
686 687
                       const char *what,
                       int timeout)
688
{
G
Guido Günther 已提交
689
    size_t got = 0;
690
    buf[0] = '\0';
691
    timeout *= 1000; /* poll wants milli seconds */
692

693
    /* Consume & discard the initial greeting */
694
    while (got < (buflen-1)) {
G
Guido Günther 已提交
695
        ssize_t ret;
696 697

        ret = read(fd, buf+got, buflen-got-1);
698

699 700 701 702 703 704
        if (ret < 0) {
            struct pollfd pfd = { .fd = fd, .events = POLLIN };
            if (errno == EINTR)
                continue;

            if (errno != EAGAIN) {
705 706 707
                virReportSystemError(conn, errno,
                                     _("Failure while reading %s startup output"),
                                     what);
708 709 710
                return -1;
            }

711
            ret = poll(&pfd, 1, timeout);
712
            if (ret == 0) {
713
                qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
714
                                 _("Timed out while reading %s startup output"), what);
715 716 717
                return -1;
            } else if (ret == -1) {
                if (errno != EINTR) {
718 719 720
                    virReportSystemError(conn, errno,
                                         _("Failure while reading %s startup output"),
                                         what);
721 722 723 724 725 726 727 728
                    return -1;
                }
            } else {
                /* Make sure we continue loop & read any further data
                   available before dealing with EOF */
                if (pfd.revents & (POLLIN | POLLHUP))
                    continue;

729
                qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
730
                                 _("Failure while reading %s startup output"), what);
731 732
                return -1;
            }
733 734
        } else if (ret == 0) {
            return 0;
735 736 737
        } else {
            got += ret;
            buf[got] = '\0';
738 739 740 741 742 743
            ret = func(conn, vm, buf, fd);
            if (ret == -1)
                return -1;
            if (ret == 1)
                continue;
            return 1;
744 745 746
        }
    }

747
    qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
748
                     _("Out of space while reading %s startup output"), what);
749 750 751 752
    return -1;

}

753 754 755 756 757 758 759 760 761

/*
 * Returns -1 for error, 0 on success
 */
static int
qemudReadLogOutput(virConnectPtr conn,
                   virDomainObjPtr vm,
                   int fd,
                   char *buf,
G
Guido Günther 已提交
762
                   size_t buflen,
763 764 765 766
                   qemudHandlerMonitorOutput func,
                   const char *what,
                   int timeout)
{
767
    int retries = (timeout*10);
768
    int got = 0;
769 770 771
    buf[0] = '\0';

    while (retries) {
772
        ssize_t func_ret, ret;
773
        int isdead = 0;
G
Guido Günther 已提交
774

775 776
        func_ret = func(conn, vm, buf, fd);

777 778
        if (kill(vm->pid, 0) == -1 && errno == ESRCH)
            isdead = 1;
779

780 781
        /* Any failures should be detected before we read the log, so we
         * always have something useful to report on failure. */
782 783
        ret = saferead(fd, buf+got, buflen-got-1);
        if (ret < 0) {
784 785 786 787 788 789
            virReportSystemError(conn, errno,
                                 _("Failure while reading %s log output"),
                                 what);
            return -1;
        }

790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805
        got += ret;
        buf[got] = '\0';
        if (got == buflen-1) {
            qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                             _("Out of space while reading %s log output"),
                             what);
            return -1;
        }

        if (isdead) {
            qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                             _("Process exited while reading %s log output"),
                             what);
            return -1;
        }

806 807
        if (func_ret <= 0)
            return func_ret;
808 809 810 811 812 813 814 815 816 817

        usleep(100*1000);
        retries--;
    }
    if (retries == 0)
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("Timed out while reading %s log output"), what);
    return -1;
}

818
static int
819
qemudCheckMonitorPrompt(virConnectPtr conn ATTRIBUTE_UNUSED,
820
                        virDomainObjPtr vm,
821 822 823 824 825 826 827 828 829 830 831
                        const char *output,
                        int fd)
{
    if (strstr(output, "(qemu) ") == NULL)
        return 1; /* keep reading */

    vm->monitor = fd;

    return 0;
}

832
static int
833 834 835 836 837
qemudOpenMonitorCommon(virConnectPtr conn,
                       struct qemud_driver* driver,
                       virDomainObjPtr vm,
                       int monfd,
                       int reconnect)
838
{
839
    char buf[1024];
840
    int ret;
841

842
    if (virSetCloseExec(monfd) < 0) {
843
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
844
                         "%s", _("Unable to set monitor close-on-exec flag"));
845
        return -1;
846
    }
847
    if (virSetNonBlock(monfd) < 0) {
848
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
849
                         "%s", _("Unable to put monitor into non-blocking mode"));
850
        return -1;
851 852
    }

853
    if (!reconnect) {
854 855 856 857
        if (qemudReadMonitorOutput(conn,
                                   vm, monfd,
                                   buf, sizeof(buf),
                                   qemudCheckMonitorPrompt,
858
                                   "monitor", 10) <= 0)
859 860 861
            ret = -1;
        else
            ret = 0;
862 863 864 865 866 867
    } else {
        vm->monitor = monfd;
        ret = 0;
    }

    if (ret != 0)
868
        return ret;
869

870 871 872
    if ((vm->monitorWatch = virEventAddHandle(vm->monitor, 0,
                                              qemudDispatchVMEvent,
                                              driver, NULL)) < 0)
873
        return -1;
874

875 876
    return 0;
}
877

878 879 880 881 882 883 884 885 886
static int
qemudOpenMonitorUnix(virConnectPtr conn,
                     struct qemud_driver* driver,
                     virDomainObjPtr vm,
                     const char *monitor,
                     int reconnect)
{
    struct sockaddr_un addr;
    int monfd;
887
    int timeout = 3; /* In seconds */
888
    int ret, i = 0;
889 890 891 892 893 894 895 896 897 898 899

    if ((monfd = socket(AF_UNIX, SOCK_STREAM, 0)) < 0) {
        virReportSystemError(conn, errno,
                             "%s", _("failed to create socket"));
        return -1;
    }

    memset(&addr, 0, sizeof(addr));
    addr.sun_family = AF_UNIX;
    strncpy(addr.sun_path, monitor, sizeof(addr.sun_path));

900 901 902 903 904 905 906 907 908 909 910 911
    do {
        ret = connect(monfd, (struct sockaddr *) &addr, sizeof(addr));

        if (ret == 0)
            break;

        if (errno == EACCES || errno == ECONNREFUSED) {
            /* EACCES       : Socket may not have shown up yet
             * ECONNREFUSED : Leftover socket hasn't been removed yet */
            continue;
        }

912 913 914
        virReportSystemError(conn, errno, "%s",
                             _("failed to connect to monitor socket"));
        goto error;
915 916 917 918 919 920 921

    } while ((++i <= timeout*5) && (usleep(.2 * 1000000) <= 0));

    if (ret != 0) {
        virReportSystemError(conn, errno, "%s",
                             _("monitor socket did not show up."));
        goto error;
922 923 924 925 926 927 928 929 930 931 932 933
    }

    if (qemudOpenMonitorCommon(conn, driver, vm, monfd, reconnect) < 0)
        goto error;

    return 0;

error:
    close(monfd);
    return -1;
}

934 935 936 937 938 939 940 941
static int
qemudOpenMonitorPty(virConnectPtr conn,
                    struct qemud_driver* driver,
                    virDomainObjPtr vm,
                    const char *monitor,
                    int reconnect)
{
    int monfd;
942

943 944 945 946 947 948 949 950 951 952 953 954
    if ((monfd = open(monitor, O_RDWR)) < 0) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("Unable to open monitor path %s"), monitor);
        return -1;
    }

    if (qemudOpenMonitorCommon(conn, driver, vm, monfd, reconnect) < 0)
        goto error;

    return 0;

error:
955
    close(monfd);
956 957 958 959 960 961 962 963 964 965
    return -1;
}

static int
qemudOpenMonitor(virConnectPtr conn,
                 struct qemud_driver *driver,
                 virDomainObjPtr vm,
                 int reconnect)
{
    switch (vm->monitor_chr->type) {
966 967 968 969
    case VIR_DOMAIN_CHR_TYPE_UNIX:
        return qemudOpenMonitorUnix(conn, driver, vm,
                                    vm->monitor_chr->data.nix.path,
                                    reconnect);
970 971 972 973 974 975 976 977 978 979
    case VIR_DOMAIN_CHR_TYPE_PTY:
        return qemudOpenMonitorPty(conn, driver, vm,
                                   vm->monitor_chr->data.file.path,
                                   reconnect);
    default:
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("unable to handle monitor type: %s"),
                         virDomainChrTypeToString(vm->monitor_chr->type));
        return -1;
    }
980 981
}

982
/* Returns -1 for error, 0 success, 1 continue reading */
983 984 985 986 987 988
static int
qemudExtractMonitorPath(virConnectPtr conn,
                        const char *haystack,
                        size_t *offset,
                        char **path)
{
989
    static const char needle[] = "char device redirected to";
990
    char *tmp, *dev;
991

992
    VIR_FREE(*path);
993
    /* First look for our magic string */
994 995 996 997 998
    if (!(tmp = strstr(haystack + *offset, needle))) {
        return 1;
    }
    tmp += sizeof(needle);
    dev = tmp;
999

1000 1001 1002 1003 1004
    /*
     * And look for first whitespace character and nul terminate
     * to mark end of the pty path
     */
    while (*tmp) {
1005
        if (c_isspace(*tmp)) {
1006
            if (VIR_ALLOC_N(*path, (tmp-dev)+1) < 0) {
1007
                virReportOOMError(conn);
1008 1009 1010 1011
                return -1;
            }
            strncpy(*path, dev, (tmp-dev));
            (*path)[(tmp-dev)] = '\0';
1012
            /* ... now further update offset till we get EOL */
1013
            *offset = tmp - haystack;
1014 1015
            return 0;
        }
1016
        tmp++;
1017 1018 1019 1020 1021
    }

    /*
     * We found a path, but didn't find any whitespace,
     * so it must be still incomplete - we should at
1022 1023
     * least see a \n - indicate that we want to carry
     * on trying again
1024
     */
1025
    return 1;
1026 1027 1028
}

static int
1029
qemudFindCharDevicePTYs(virConnectPtr conn,
1030
                        virDomainObjPtr vm,
1031 1032
                        const char *output,
                        int fd ATTRIBUTE_UNUSED)
1033
{
1034
    size_t offset = 0;
1035
    int ret, i;
1036 1037

    /* The order in which QEMU prints out the PTY paths is
1038 1039
       the order in which it procsses its serial and parallel
       device args. This code must match that ordering.... */
1040

1041
    /* first comes the serial devices */
1042 1043
    for (i = 0 ; i < vm->def->nserials ; i++) {
        virDomainChrDefPtr chr = vm->def->serials[i];
1044 1045 1046
        if (chr->type == VIR_DOMAIN_CHR_TYPE_PTY) {
            if ((ret = qemudExtractMonitorPath(conn, output, &offset,
                                               &chr->data.file.path)) != 0)
1047
                return ret;
1048 1049 1050
        }
    }

1051
    /* then the parallel devices */
1052 1053
    for (i = 0 ; i < vm->def->nparallels ; i++) {
        virDomainChrDefPtr chr = vm->def->parallels[i];
1054 1055 1056
        if (chr->type == VIR_DOMAIN_CHR_TYPE_PTY) {
            if ((ret = qemudExtractMonitorPath(conn, output, &offset,
                                               &chr->data.file.path)) != 0)
1057
                return ret;
1058 1059 1060
        }
    }

1061
    return 0;
1062 1063
}

1064 1065 1066 1067
static int
qemudWaitForMonitor(virConnectPtr conn,
                    struct qemud_driver* driver,
                    virDomainObjPtr vm, off_t pos)
1068
{
1069
    char buf[4096]; /* Plenty of space to get startup greeting */
1070 1071 1072 1073 1074
    int logfd;
    int ret;

    if ((logfd = qemudLogReadFD(conn, driver->logDir, vm->def->name, pos))
        < 0)
1075
        return -1;
1076

1077 1078 1079
    ret = qemudReadLogOutput(conn, vm, logfd, buf, sizeof(buf),
                             qemudFindCharDevicePTYs,
                             "console", 3);
1080
    if (close(logfd) < 0) {
1081
        char ebuf[4096];
1082
        VIR_WARN(_("Unable to close logfile: %s\n"),
1083 1084
                 virStrerror(errno, ebuf, sizeof ebuf));
    }
1085

1086 1087 1088 1089 1090 1091
    if (ret < 0) {
        /* Unexpected end of file - inform user of QEMU log data */
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("unable to start guest: %s"), buf);
        return -1;
    }
1092

1093 1094 1095 1096
    if (qemudOpenMonitor(conn, driver, vm, 0) < 0)
        return -1;

    return 0;
1097 1098
}

1099 1100
static int
qemudDetectVcpuPIDs(virConnectPtr conn,
1101
                    virDomainObjPtr vm) {
1102 1103 1104 1105 1106 1107
    char *qemucpus = NULL;
    char *line;
    int lastVcpu = -1;

    /* Only KVM has seperate threads for CPUs,
       others just use main QEMU process for CPU */
1108
    if (vm->def->virtType != VIR_DOMAIN_VIRT_KVM)
1109 1110 1111 1112 1113
        vm->nvcpupids = 1;
    else
        vm->nvcpupids = vm->def->vcpus;

    if (VIR_ALLOC_N(vm->vcpupids, vm->nvcpupids) < 0) {
1114
        virReportOOMError(conn);
1115 1116 1117
        return -1;
    }

1118
    if (vm->def->virtType != VIR_DOMAIN_VIRT_KVM) {
1119 1120 1121 1122
        vm->vcpupids[0] = vm->pid;
        return 0;
    }

1123
    if (qemudMonitorCommand(vm, "info cpus", &qemucpus) < 0) {
1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                         "%s", _("cannot run monitor command to fetch CPU thread info"));
        VIR_FREE(vm->vcpupids);
        vm->nvcpupids = 0;
        return -1;
    }

    /*
     * This is the gross format we're about to parse :-{
     *
     * (qemu) info cpus
     * * CPU #0: pc=0x00000000000f0c4a thread_id=30019
     *   CPU #1: pc=0x00000000fffffff0 thread_id=30020
     *   CPU #2: pc=0x00000000fffffff0 thread_id=30021
     *
     */
    line = qemucpus;
    do {
        char *offset = strchr(line, '#');
        char *end = NULL;
        int vcpu = 0, tid = 0;

        /* See if we're all done */
        if (offset == NULL)
            break;

        /* Extract VCPU number */
        if (virStrToLong_i(offset + 1, &end, 10, &vcpu) < 0)
            goto error;
        if (end == NULL || *end != ':')
            goto error;

        /* Extract host Thread ID */
        if ((offset = strstr(line, "thread_id=")) == NULL)
            goto error;
        if (virStrToLong_i(offset + strlen("thread_id="), &end, 10, &tid) < 0)
            goto error;
        if (end == NULL || !c_isspace(*end))
            goto error;

        /* Validate the VCPU is in expected range & order */
        if (vcpu > vm->nvcpupids ||
            vcpu != (lastVcpu + 1))
            goto error;

        lastVcpu = vcpu;
        vm->vcpupids[vcpu] = tid;

        /* Skip to next data line */
        line = strchr(offset, '\r');
        if (line == NULL)
            line = strchr(offset, '\n');
    } while (line != NULL);

    /* Validate we got data for all VCPUs we expected */
    if (lastVcpu != (vm->def->vcpus - 1))
        goto error;

1182
    VIR_FREE(qemucpus);
1183 1184 1185 1186
    return 0;

error:
    VIR_FREE(vm->vcpupids);
1187 1188
    vm->nvcpupids = 0;
    VIR_FREE(qemucpus);
1189 1190 1191 1192 1193 1194 1195 1196

    /* Explicitly return success, not error. Older KVM does
       not have vCPU -> Thread mapping info and we don't
       want to break its use. This merely disables ability
       to pin vCPUS with libvirt */
    return 0;
}

1197 1198
static int
qemudInitCpus(virConnectPtr conn,
D
Daniel Veillard 已提交
1199 1200
              virDomainObjPtr vm,
              const char *migrateFrom) {
1201 1202 1203 1204 1205 1206
    char *info = NULL;
#if HAVE_SCHED_GETAFFINITY
    cpu_set_t mask;
    int i, maxcpu = QEMUD_CPUMASK_LEN;
    virNodeInfo nodeinfo;

1207
    if (nodeGetInfo(conn, &nodeinfo) < 0)
1208 1209 1210 1211 1212 1213 1214 1215
        return -1;

    /* setaffinity fails if you set bits for CPUs which
     * aren't present, so we have to limit ourselves */
    if (maxcpu > nodeinfo.cpus)
        maxcpu = nodeinfo.cpus;

    CPU_ZERO(&mask);
D
Daniel P. Berrange 已提交
1216 1217 1218 1219 1220 1221
    if (vm->def->cpumask) {
        for (i = 0 ; i < maxcpu ; i++)
            if (vm->def->cpumask[i])
                CPU_SET(i, &mask);
    } else {
        for (i = 0 ; i < maxcpu ; i++)
1222
            CPU_SET(i, &mask);
D
Daniel P. Berrange 已提交
1223
    }
1224 1225 1226 1227

    for (i = 0 ; i < vm->nvcpupids ; i++) {
        if (sched_setaffinity(vm->vcpupids[i],
                              sizeof(mask), &mask) < 0) {
1228 1229
            virReportSystemError(conn, errno, "%s",
                                 _("failed to set CPU affinity"));
1230 1231 1232 1233 1234
            return -1;
        }
    }
#endif /* HAVE_SCHED_GETAFFINITY */

D
Daniel Veillard 已提交
1235 1236
    if (migrateFrom == NULL) {
        /* Allow the CPUS to start executing */
1237
        if (qemudMonitorCommand(vm, "cont", &info) < 0) {
D
Daniel Veillard 已提交
1238 1239 1240 1241 1242
            qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                             "%s", _("resume operation failed"));
            return -1;
        }
        VIR_FREE(info);
1243 1244 1245 1246 1247 1248
    }

    return 0;
}


1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260
static int
qemudInitPasswords(virConnectPtr conn,
                   struct qemud_driver *driver,
                   virDomainObjPtr vm) {
    char *info = NULL;

    /*
     * NB: Might have more passwords to set in the future. eg a qcow
     * disk decryption password, but there's no monitor command
     * for that yet...
     */

1261 1262 1263
    if ((vm->def->ngraphics == 1) &&
        vm->def->graphics[0]->type == VIR_DOMAIN_GRAPHICS_TYPE_VNC &&
        (vm->def->graphics[0]->data.vnc.passwd || driver->vncPassword)) {
1264 1265

        if (qemudMonitorCommandExtra(vm, "change vnc password",
1266 1267
                                     vm->def->graphics[0]->data.vnc.passwd ?
                                     vm->def->graphics[0]->data.vnc.passwd :
1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281
                                     driver->vncPassword,
                                     QEMU_PASSWD_PROMPT,
                                     &info) < 0) {
            qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                             "%s", _("setting VNC password failed"));
            return -1;
        }
        VIR_FREE(info);
    }

    return 0;
}


1282
static int qemudNextFreeVNCPort(struct qemud_driver *driver ATTRIBUTE_UNUSED) {
1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317
    int i;

    for (i = 5900 ; i < 6000 ; i++) {
        int fd;
        int reuse = 1;
        struct sockaddr_in addr;
        addr.sin_family = AF_INET;
        addr.sin_port = htons(i);
        addr.sin_addr.s_addr = htonl(INADDR_ANY);
        fd = socket(PF_INET, SOCK_STREAM, 0);
        if (fd < 0)
            return -1;

        if (setsockopt(fd, SOL_SOCKET, SO_REUSEADDR, (void*)&reuse, sizeof(reuse)) < 0) {
            close(fd);
            break;
        }

        if (bind(fd, (struct sockaddr*)&addr, sizeof(addr)) == 0) {
            /* Not in use, lets grab it */
            close(fd);
            return i;
        }
        close(fd);

        if (errno == EADDRINUSE) {
            /* In use, try next */
            continue;
        }
        /* Some other bad failure, get out.. */
        break;
    }
    return -1;
}

1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335
static int qemuPrepareHostDevices(virConnectPtr conn,
                                  virDomainDefPtr def) {
    int i;

    /* We have to use 2 loops here. *All* devices must
     * be detached before we reset any of them, because
     * in some cases you have to reset the whole PCI,
     * which impacts all devices on it
     */

    for (i = 0 ; i < def->nhostdevs ; i++) {
        virDomainHostdevDefPtr hostdev = def->hostdevs[i];

        if (hostdev->mode != VIR_DOMAIN_HOSTDEV_MODE_SUBSYS)
            continue;
        if (hostdev->source.subsys.type != VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_PCI)
            continue;

1336
        if (hostdev->managed) {
1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390
            pciDevice *dev = pciGetDevice(conn,
                                          hostdev->source.subsys.u.pci.domain,
                                          hostdev->source.subsys.u.pci.bus,
                                          hostdev->source.subsys.u.pci.slot,
                                          hostdev->source.subsys.u.pci.function);
            if (!dev)
                goto error;

            if (pciDettachDevice(conn, dev) < 0) {
                pciFreeDevice(conn, dev);
                goto error;
            }

            pciFreeDevice(conn, dev);
        } /* else {
             XXX validate that non-managed device isn't in use, eg
             by checking that device is either un-bound, or bound
             to pci-stub.ko
        } */
    }

    /* Now that all the PCI hostdevs have be dettached, we can safely
     * reset them */
    for (i = 0 ; i < def->nhostdevs ; i++) {
        virDomainHostdevDefPtr hostdev = def->hostdevs[i];
        pciDevice *dev;

        if (hostdev->mode != VIR_DOMAIN_HOSTDEV_MODE_SUBSYS)
            continue;
        if (hostdev->source.subsys.type != VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_PCI)
            continue;

        dev = pciGetDevice(conn,
                           hostdev->source.subsys.u.pci.domain,
                           hostdev->source.subsys.u.pci.bus,
                           hostdev->source.subsys.u.pci.slot,
                           hostdev->source.subsys.u.pci.function);
        if (!dev)
            goto error;

        if (pciResetDevice(conn, dev) < 0) {
            pciFreeDevice(conn, dev);
            goto error;
        }

        pciFreeDevice(conn, dev);
    }

    return 0;

error:
    return -1;
}

1391 1392 1393 1394 1395 1396 1397 1398 1399 1400
static const char *const defaultDeviceACL[] = {
    "/dev/null", "/dev/full", "/dev/zero",
    "/dev/random", "/dev/urandom",
    "/dev/ptmx", "/dev/kvm", "/dev/kqemu",
    "/dev/rtc", "/dev/hpet", "/dev/net/tun",
    NULL,
};
#define DEVICE_PTY_MAJOR 136
#define DEVICE_SND_MAJOR 116

1401 1402 1403 1404 1405 1406
static int qemuSetupCgroup(virConnectPtr conn,
                           struct qemud_driver *driver,
                           virDomainObjPtr vm)
{
    virCgroupPtr cgroup = NULL;
    int rc;
1407 1408
    unsigned int i;
    const char *const *deviceACL = defaultDeviceACL;
1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420

    if (driver->cgroup == NULL)
        return 0; /* Not supported, so claim success */

    rc = virCgroupForDomain(driver->cgroup, vm->def->name, &cgroup, 1);
    if (rc != 0) {
        virReportSystemError(conn, -rc,
                             _("Unable to create cgroup for %s"),
                             vm->def->name);
        goto cleanup;
    }

1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476
    rc = virCgroupDenyAllDevices(cgroup);
    if (rc != 0) {
        if (rc == -EPERM) {
            VIR_WARN0("Group devices ACL is not accessible, disabling whitelisting");
            goto done;
        }

        virReportSystemError(conn, -rc,
                             _("Unable to deny all devices for %s"), vm->def->name);
        goto cleanup;
    }

    for (i = 0; i < vm->def->ndisks ; i++) {
        if (vm->def->disks[i]->type != VIR_DOMAIN_DISK_TYPE_BLOCK ||
            vm->def->disks[i]->src == NULL)
            continue;

        rc = virCgroupAllowDevicePath(cgroup,
                                      vm->def->disks[i]->src);
        if (rc != 0) {
            virReportSystemError(conn, -rc,
                                 _("Unable to allow device %s for %s"),
                                 vm->def->disks[i]->src, vm->def->name);
            goto cleanup;
        }
    }

    rc = virCgroupAllowDeviceMajor(cgroup, 'c', DEVICE_PTY_MAJOR);
    if (rc != 0) {
        virReportSystemError(conn, -rc, "%s",
                             _("unable to allow /dev/pts/ devices"));
        goto cleanup;
    }

    if (vm->def->nsounds) {
        rc = virCgroupAllowDeviceMajor(cgroup, 'c', DEVICE_SND_MAJOR);
        if (rc != 0) {
            virReportSystemError(conn, -rc, "%s",
                                 _("unable to allow /dev/snd/ devices"));
            goto cleanup;
        }
    }

    for (i = 0; deviceACL[i] != NULL ; i++) {
        rc = virCgroupAllowDevicePath(cgroup,
                                      deviceACL[i]);
        if (rc < 0 &&
            rc != -ENOENT) {
            virReportSystemError(conn, -rc,
                                 _("unable to allow device %s"),
                                 deviceACL[i]);
            goto cleanup;
        }
    }

done:
1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545
    virCgroupFree(&cgroup);
    return 0;

cleanup:
    if (cgroup) {
        virCgroupRemove(cgroup);
        virCgroupFree(&cgroup);
    }
    return -1;
}


static int qemuRemoveCgroup(virConnectPtr conn,
                            struct qemud_driver *driver,
                            virDomainObjPtr vm)
{
    virCgroupPtr cgroup;
    int rc;

    if (driver->cgroup == NULL)
        return 0; /* Not supported, so claim success */

    rc = virCgroupForDomain(driver->cgroup, vm->def->name, &cgroup, 0);
    if (rc != 0) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("Unable to find cgroup for %s\n"),
                         vm->def->name);
        return rc;
    }

    rc = virCgroupRemove(cgroup);
    virCgroupFree(&cgroup);
    return rc;
}

static int qemuAddToCgroup(struct qemud_driver *driver,
                           virDomainDefPtr def)
{
    virCgroupPtr cgroup = NULL;
    int ret = -1;
    int rc;

    if (driver->cgroup == NULL)
        return 0; /* Not supported, so claim success */

    rc = virCgroupForDomain(driver->cgroup, def->name, &cgroup, 0);
    if (rc != 0) {
        virReportSystemError(NULL, -rc,
                             _("unable to find cgroup for domain %s"),
                             def->name);
        goto cleanup;
    }

    rc = virCgroupAddTask(cgroup, getpid());
    if (rc != 0) {
        virReportSystemError(NULL, -rc,
                             _("unable to add domain %s task %d to cgroup"),
                             def->name, getpid());
        goto cleanup;
    }

    ret = 0;

cleanup:
    virCgroupFree(&cgroup);
    return ret;
}


1546 1547 1548 1549 1550 1551 1552 1553 1554
static int qemudDomainSetSecurityLabel(virConnectPtr conn, struct qemud_driver *driver, virDomainObjPtr vm)
{
    if (vm->def->seclabel.label != NULL)
        if (driver->securityDriver && driver->securityDriver->domainSetSecurityLabel)
            return driver->securityDriver->domainSetSecurityLabel(conn, driver->securityDriver,
                                                                 vm);
    return 0;
}

1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753

#ifdef __linux__
static int qemuDomainSetHostdevUSBOwnership(virConnectPtr conn,
                                            virDomainHostdevDefPtr def,
                                            uid_t uid, gid_t gid)
{
    char *usbpath = NULL;

    /* XXX what todo for USB devs assigned based on product/vendor ? Doom :-( */
    if (!def->source.subsys.u.usb.bus ||
        !def->source.subsys.u.usb.device)
        return 0;

    if (virAsprintf(&usbpath, "/dev/bus/usb/%03d/%03d",
                    def->source.subsys.u.usb.bus,
                    def->source.subsys.u.usb.device) < 0) {
        virReportOOMError(conn);
        return -1;
    }

    VIR_DEBUG("Setting ownership on %s to %d:%d", usbpath, uid, gid);
    if (chown(usbpath, uid, gid) < 0) {
        virReportSystemError(conn, errno, _("cannot set ownership on %s"), usbpath);
        VIR_FREE(usbpath);
        return -1;
    }
    VIR_FREE(usbpath);

    return 0;
}

static int qemuDomainSetHostdevPCIOwnership(virConnectPtr conn,
                                            virDomainHostdevDefPtr def,
                                            uid_t uid, gid_t gid)
{
    char *pcidir = NULL;
    char *file = NULL;
    DIR *dir = NULL;
    int ret = -1;
    struct dirent *ent;

    if (virAsprintf(&pcidir, "/sys/bus/pci/devices/%04x:%02x:%02x.%x",
                    def->source.subsys.u.pci.domain,
                    def->source.subsys.u.pci.bus,
                    def->source.subsys.u.pci.slot,
                    def->source.subsys.u.pci.function) < 0) {
        virReportOOMError(conn);
        goto cleanup;
    }

    if (!(dir = opendir(pcidir))) {
        virReportSystemError(conn, errno,
                             _("cannot open %s"), pcidir);
        goto cleanup;
    }

    while ((ent = readdir(dir)) != NULL) {
        /* QEMU device assignment requires:
         *   $PCIDIR/config, $PCIDIR/resource, $PCIDIR/resourceNNN, $PCIDIR/rom
         */
        if (STREQ(ent->d_name, "config") ||
            STRPREFIX(ent->d_name, "resource") ||
            STREQ(ent->d_name, "rom")) {
            if (virAsprintf(&file, "%s/%s", pcidir, ent->d_name) < 0) {
                virReportOOMError(conn);
                goto cleanup;
            }
            VIR_DEBUG("Setting ownership on %s to %d:%d", file, uid, gid);
            if (chown(file, uid, gid) < 0) {
                virReportSystemError(conn, errno, _("cannot set ownership on %s"), file);
                goto cleanup;
            }
            VIR_FREE(file);
        }
    }

    ret = 0;

cleanup:
    if (dir)
        closedir(dir);
    VIR_FREE(file);
    VIR_FREE(pcidir);
    return ret;
}
#endif


static int qemuDomainSetHostdevOwnership(virConnectPtr conn,
                                         virDomainHostdevDefPtr def,
                                         uid_t uid, gid_t gid)
{
    if (def->mode != VIR_DOMAIN_HOSTDEV_MODE_SUBSYS)
        return 0;

#ifdef __linux__
    switch (def->source.subsys.type) {
    case VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_USB:
        return qemuDomainSetHostdevUSBOwnership(conn, def, uid, gid);

    case VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_PCI:
        return qemuDomainSetHostdevPCIOwnership(conn, def, uid, gid);

    }
    return 0;
#else
    qemudReportError(conn, NULL, NULL, "%s",
                     _("unable to set host device ownership on this platform"));
    return -1;
#endif

}

static int qemuDomainSetDiskOwnership(virConnectPtr conn,
                                      virDomainDiskDefPtr def,
                                      uid_t uid, gid_t gid)
{

    if (!def->src)
        return 0;

    VIR_DEBUG("Setting ownership on %s to %d:%d", def->src, uid, gid);
    if (chown(def->src, uid, gid) < 0) {
        virReportSystemError(conn, errno, _("cannot set ownership on %s"),
                             def->src);
        return -1;
    }
    return 0;
}

static int qemuDomainSetDeviceOwnership(virConnectPtr conn,
                                        struct qemud_driver *driver,
                                        virDomainDeviceDefPtr def,
                                        int restore)
{
    uid_t uid;
    gid_t gid;

    if (!driver->privileged)
        return 0;

    /* short circuit case of root:root */
    if (!driver->user && !driver->group)
        return 0;

    uid = restore ? 0 : driver->user;
    gid = restore ? 0 : driver->group;

    switch (def->type) {
    case VIR_DOMAIN_DEVICE_DISK:
        if (restore &&
            (def->data.disk->readonly || def->data.disk->shared))
            return 0;

        return qemuDomainSetDiskOwnership(conn, def->data.disk, uid, gid);

    case VIR_DOMAIN_DEVICE_HOSTDEV:
        return qemuDomainSetHostdevOwnership(conn, def->data.hostdev, uid, gid);
    }

    return 0;
}

static int qemuDomainSetAllDeviceOwnership(virConnectPtr conn,
                                           struct qemud_driver *driver,
                                           virDomainDefPtr def,
                                           int restore)
{
    int i;
    uid_t uid;
    gid_t gid;

    if (!driver->privileged)
        return 0;

    /* short circuit case of root:root */
    if (!driver->user && !driver->group)
        return 0;

    uid = restore ? 0 : driver->user;
    gid = restore ? 0 : driver->group;

    for (i = 0 ; i < def->ndisks ; i++) {
        if (restore &&
            (def->disks[i]->readonly || def->disks[i]->shared))
            continue;

        if (qemuDomainSetDiskOwnership(conn, def->disks[i], uid, gid) < 0)
            return -1;
    }

    for (i = 0 ; i < def->nhostdevs ; i++) {
        if (qemuDomainSetHostdevOwnership(conn, def->hostdevs[i], uid, gid) < 0)
            return -1;
    }

    return 0;
}

1754 1755 1756
static virDomainPtr qemudDomainLookupByName(virConnectPtr conn,
                                            const char *name);

1757 1758 1759 1760
struct qemudHookData {
    virConnectPtr conn;
    virDomainObjPtr vm;
    struct qemud_driver *driver;
1761 1762 1763
};

static int qemudSecurityHook(void *data) {
1764 1765 1766 1767
    struct qemudHookData *h = data;

    if (qemuAddToCgroup(h->driver, h->vm->def) < 0)
        return -1;
1768 1769 1770

    if (qemudDomainSetSecurityLabel(h->conn, h->driver, h->vm) < 0) {
        qemudReportError(h->conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
1771
                         "%s", _("Failed to set security label"));
1772 1773 1774 1775 1776 1777 1778 1779
        return -1;
    }

    if (h->driver->privileged) {
        DEBUG("Dropping privileges of VM to %d:%d", h->driver->user, h->driver->group);

        if (qemuDomainSetAllDeviceOwnership(h->conn, h->driver, h->vm->def, 0) < 0)
            return -1;
1780

1781 1782 1783 1784 1785
        if (h->driver->group) {
            if (setregid(h->driver->group, h->driver->group) < 0) {
                virReportSystemError(NULL, errno,
                                     _("cannot change to '%d' group"),
                                     h->driver->group);
1786
                return -1;
1787
            }
1788
        }
1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799
        if (h->driver->user) {
            if (setreuid(h->driver->user, h->driver->user) < 0) {
                virReportSystemError(NULL, errno,
                                     _("cannot change to '%d' user"),
                                     h->driver->user);
                return -1;
            }
        }
    }

    return 0;
1800 1801
}

1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819
static int
qemuPrepareMonitorChr(virConnectPtr conn,
                      struct qemud_driver *driver,
                      virDomainChrDefPtr monitor_chr,
                      const char *vm)
{
    monitor_chr->type = VIR_DOMAIN_CHR_TYPE_UNIX;
    monitor_chr->data.nix.listen = 1;

    if (virAsprintf(&monitor_chr->data.nix.path, "%s/%s.monitor",
                    driver->stateDir, vm) < 0) {
        virReportOOMError(conn);
        return -1;
    }

    return 0;
}

1820 1821
static int qemudStartVMDaemon(virConnectPtr conn,
                              struct qemud_driver *driver,
1822
                              virDomainObjPtr vm,
1823 1824
                              const char *migrateFrom,
                              int stdin_fd) {
1825
    const char **argv = NULL, **tmp;
1826
    const char **progenv = NULL;
1827
    int i, ret;
1828
    struct stat sb;
1829 1830
    int *tapfds = NULL;
    int ntapfds = 0;
1831
    unsigned int qemuCmdFlags;
1832
    fd_set keepfd;
1833
    const char *emulator;
G
Guido Günther 已提交
1834
    pid_t child;
1835
    int pos = -1;
1836
    char ebuf[1024];
1837
    char *pidfile = NULL;
1838
    int logfile;
1839

1840
    struct qemudHookData hookData;
1841 1842 1843 1844
    hookData.conn = conn;
    hookData.vm = vm;
    hookData.driver = driver;

1845
    FD_ZERO(&keepfd);
1846

1847
    if (virDomainIsActive(vm)) {
1848
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_INVALID,
1849
                         "%s", _("VM is already active"));
1850 1851 1852
        return -1;
    }

1853 1854 1855 1856 1857 1858 1859 1860
    /* If you are using a SecurityDriver with dynamic labelling,
       then generate a security label for isolation */
    if (vm->def->seclabel.type == VIR_DOMAIN_SECLABEL_DYNAMIC &&
        driver->securityDriver &&
        driver->securityDriver->domainGenSecurityLabel &&
        driver->securityDriver->domainGenSecurityLabel(conn, vm) < 0)
        return -1;

1861 1862 1863
    /* Ensure no historical cgroup for this VM is lieing around bogus settings */
    qemuRemoveCgroup(conn, driver, vm);

1864 1865 1866
    if ((vm->def->ngraphics == 1) &&
        vm->def->graphics[0]->type == VIR_DOMAIN_GRAPHICS_TYPE_VNC &&
        vm->def->graphics[0]->data.vnc.autoport) {
1867
        int port = qemudNextFreeVNCPort(driver);
1868
        if (port < 0) {
1869
            qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
1870
                             "%s", _("Unable to find an unused VNC port"));
1871
            goto cleanup;
1872
        }
1873
        vm->def->graphics[0]->data.vnc.port = port;
1874
    }
1875

1876
    if (virFileMakePath(driver->logDir) < 0) {
1877 1878 1879
        virReportSystemError(conn, errno,
                             _("cannot create log directory %s"),
                             driver->logDir);
1880
        goto cleanup;
1881 1882
    }

1883
    if ((logfile = qemudLogFD(conn, driver, vm->def->name)) < 0)
1884
        goto cleanup;
1885

1886 1887
    emulator = vm->def->emulator;

1888 1889 1890 1891
    /* Make sure the binary we are about to try exec'ing exists.
     * Technically we could catch the exec() failure, but that's
     * in a sub-process so its hard to feed back a useful error
     */
1892
    if (stat(emulator, &sb) < 0) {
1893 1894 1895
        virReportSystemError(conn, errno,
                             _("Cannot find QEMU binary %s"),
                             emulator);
1896
        goto cleanup;
1897 1898
    }

1899
    if (qemudExtractVersionInfo(emulator,
1900
                                NULL,
1901
                                &qemuCmdFlags) < 0)
1902
        goto cleanup;
1903

1904 1905 1906
    if (qemuSetupCgroup(conn, driver, vm) < 0)
        goto cleanup;

1907
    if (qemuPrepareHostDevices(conn, vm->def) < 0)
1908
        goto cleanup;
1909

1910 1911 1912 1913 1914
    if (VIR_ALLOC(vm->monitor_chr) < 0) {
        virReportOOMError(conn);
        goto cleanup;
    }

1915 1916
    if (qemuPrepareMonitorChr(conn, driver, vm->monitor_chr, vm->def->name) < 0)
        goto cleanup;
1917

D
Daniel P. Berrange 已提交
1918 1919 1920 1921 1922 1923 1924
    if ((ret = virFileDeletePid(driver->stateDir, vm->def->name)) != 0) {
        virReportSystemError(conn, ret,
                             _("Cannot remove stale PID file for %s"),
                             vm->def->name);
        goto cleanup;
    }

1925 1926 1927 1928 1929
    if (!(pidfile = virFilePid(driver->stateDir, vm->def->name))) {
        virReportSystemError(conn, errno,
                             "%s", _("Failed to build pidfile path."));
        goto cleanup;
    }
D
Daniel P. Berrange 已提交
1930

1931
    vm->def->id = driver->nextvmid++;
1932
    if (qemudBuildCommandLine(conn, driver, vm->def, vm->monitor_chr,
1933
                              qemuCmdFlags, &argv, &progenv,
1934 1935
                              &tapfds, &ntapfds, migrateFrom) < 0)
        goto cleanup;
1936

1937 1938
    tmp = progenv;
    while (*tmp) {
1939
        if (safewrite(logfile, *tmp, strlen(*tmp)) < 0)
1940
            VIR_WARN(_("Unable to write envv to logfile: %s\n"),
1941
                     virStrerror(errno, ebuf, sizeof ebuf));
1942
        if (safewrite(logfile, " ", 1) < 0)
1943
            VIR_WARN(_("Unable to write envv to logfile: %s\n"),
1944
                     virStrerror(errno, ebuf, sizeof ebuf));
1945 1946
        tmp++;
    }
1947 1948
    tmp = argv;
    while (*tmp) {
1949
        if (safewrite(logfile, *tmp, strlen(*tmp)) < 0)
1950
            VIR_WARN(_("Unable to write argv to logfile: %s\n"),
1951
                     virStrerror(errno, ebuf, sizeof ebuf));
1952
        if (safewrite(logfile, " ", 1) < 0)
1953
            VIR_WARN(_("Unable to write argv to logfile: %s\n"),
1954
                     virStrerror(errno, ebuf, sizeof ebuf));
1955 1956
        tmp++;
    }
1957
    if (safewrite(logfile, "\n", 1) < 0)
1958
        VIR_WARN(_("Unable to write argv to logfile: %s\n"),
1959
                 virStrerror(errno, ebuf, sizeof ebuf));
1960

1961
    if ((pos = lseek(logfile, 0, SEEK_END)) < 0)
1962
        VIR_WARN(_("Unable to seek to end of logfile: %s\n"),
1963
                 virStrerror(errno, ebuf, sizeof ebuf));
1964

1965 1966 1967
    for (i = 0 ; i < ntapfds ; i++)
        FD_SET(tapfds[i], &keepfd);

1968
    ret = virExecDaemonize(conn, argv, progenv, &keepfd, &child,
1969
                           stdin_fd, &logfile, &logfile,
1970
                           VIR_EXEC_NONBLOCK | VIR_EXEC_CLEAR_CAPS,
1971 1972 1973
                           qemudSecurityHook, &hookData,
                           pidfile);
    VIR_FREE(pidfile);
G
Guido Günther 已提交
1974 1975 1976

    /* wait for qemu process to to show up */
    if (ret == 0) {
1977
        if (virFileReadPid(driver->stateDir, vm->def->name, &vm->pid)) {
1978
            qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
1979
                             _("Domain %s didn't show up\n"), vm->def->name);
1980
            ret = -1;
G
Guido Günther 已提交
1981
        }
1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992
    } else if (ret == -2) {
        /* The virExec process that launches the daemon failed. Pending on
         * when it failed (we can't determine for sure), there may be
         * extra info in the domain log (if the hook failed for example).
         *
         * Pretend like things succeeded, and let 'WaitForMonitor' report
         * the log contents for us.
         */
        vm->pid = child;
        ret = 0;
    }
1993 1994

    vm->state = migrateFrom ? VIR_DOMAIN_PAUSED : VIR_DOMAIN_RUNNING;
1995

1996
    for (i = 0 ; argv[i] ; i++)
1997 1998
        VIR_FREE(argv[i]);
    VIR_FREE(argv);
1999

2000 2001 2002 2003
    for (i = 0 ; progenv[i] ; i++)
        VIR_FREE(progenv[i]);
    VIR_FREE(progenv);

2004 2005 2006
    if (tapfds) {
        for (i = 0 ; i < ntapfds ; i++) {
            close(tapfds[i]);
2007
        }
2008
        VIR_FREE(tapfds);
2009 2010
    }

2011 2012 2013 2014 2015 2016 2017 2018
    if (ret == -1)
        goto cleanup;

    if ((qemudWaitForMonitor(conn, driver, vm, pos) < 0) ||
        (qemudDetectVcpuPIDs(conn, vm) < 0) ||
        (qemudInitCpus(conn, vm, migrateFrom) < 0) ||
        (qemudInitPasswords(conn, driver, vm) < 0) ||
        (qemudDomainSetMemoryBalloon(conn, vm, vm->def->memory) < 0) ||
2019
        (virDomainSaveStatus(conn, driver->stateDir, vm) < 0)) {
2020 2021 2022
        qemudShutdownVMDaemon(conn, driver, vm);
        ret = -1;
        /* No need for 'goto cleanup' now since qemudShutdownVMDaemon does enough */
2023 2024
    }

2025
    return ret;
2026 2027 2028 2029 2030 2031 2032

cleanup:
    if (vm->def->seclabel.type == VIR_DOMAIN_SECLABEL_DYNAMIC) {
        VIR_FREE(vm->def->seclabel.model);
        VIR_FREE(vm->def->seclabel.label);
        VIR_FREE(vm->def->seclabel.imagelabel);
    }
2033
    qemuRemoveCgroup(conn, driver, vm);
2034 2035 2036 2037
    if ((vm->def->ngraphics == 1) &&
        vm->def->graphics[0]->type == VIR_DOMAIN_GRAPHICS_TYPE_VNC &&
        vm->def->graphics[0]->data.vnc.autoport)
        vm->def->graphics[0]->data.vnc.port = -1;
2038 2039
    if (logfile != -1)
        close(logfile);
2040 2041
    vm->def->id = -1;
    return -1;
2042 2043 2044
}


2045
static void qemudShutdownVMDaemon(virConnectPtr conn,
2046 2047
                                  struct qemud_driver *driver,
                                  virDomainObjPtr vm) {
D
Daniel P. Berrange 已提交
2048
    int ret;
2049
    int retries = 0;
D
Daniel P. Berrange 已提交
2050

2051
    if (!virDomainIsActive(vm))
2052
        return;
2053

2054
    VIR_DEBUG(_("Shutting down VM '%s'\n"), vm->def->name);
2055

G
Guido Günther 已提交
2056 2057
    if (virKillProcess(vm->pid, 0) == 0 &&
        virKillProcess(vm->pid, SIGTERM) < 0)
2058 2059 2060
        virReportSystemError(conn, errno,
                             _("Failed to send SIGTERM to %s (%d)"),
                             vm->def->name, vm->pid);
2061

2062 2063 2064
    if (vm->monitorWatch != -1) {
        virEventRemoveHandle(vm->monitorWatch);
        vm->monitorWatch = -1;
2065
    }
2066 2067 2068 2069 2070

    if (vm->monitor != -1)
        close(vm->monitor);
    vm->monitor = -1;

G
Guido Günther 已提交
2071 2072
    /* shut it off for sure */
    virKillProcess(vm->pid, SIGKILL);
2073

2074 2075 2076 2077
    /* Reset Security Labels */
    if (driver->securityDriver)
        driver->securityDriver->domainRestoreSecurityLabel(conn, vm);

2078 2079 2080 2081 2082 2083 2084
    /* Clear out dynamically assigned labels */
    if (vm->def->seclabel.type == VIR_DOMAIN_SECLABEL_DYNAMIC) {
        VIR_FREE(vm->def->seclabel.model);
        VIR_FREE(vm->def->seclabel.label);
        VIR_FREE(vm->def->seclabel.imagelabel);
    }

2085 2086 2087 2088
    if (qemuDomainSetAllDeviceOwnership(conn, driver, vm->def, 1) < 0)
        VIR_WARN("Failed to restore all device ownership for %s",
                 vm->def->name);

2089 2090 2091 2092 2093 2094 2095 2096 2097 2098
retry:
    if ((ret = qemuRemoveCgroup(conn, driver, vm)) < 0) {
        if (ret == -EBUSY && (retries++ < 5)) {
            usleep(200*1000);
            goto retry;
        }
        VIR_WARN("Failed to remove cgroup for %s",
                 vm->def->name);
    }

2099
    if (qemudRemoveDomainStatus(conn, driver, vm) < 0) {
2100
        VIR_WARN(_("Failed to remove domain status for %s"),
2101 2102
                 vm->def->name);
    }
D
Daniel P. Berrange 已提交
2103 2104 2105 2106 2107 2108
    if ((ret = virFileDeletePid(driver->stateDir, vm->def->name)) != 0) {
        char ebuf[1024];
        VIR_WARN(_("Failed to remove PID file for %s: %s"),
                 vm->def->name, virStrerror(errno, ebuf, sizeof ebuf));
    }

2109
    vm->pid = -1;
2110
    vm->def->id = -1;
2111
    vm->state = VIR_DOMAIN_SHUTOFF;
2112
    VIR_FREE(vm->vcpupids);
2113
    vm->nvcpupids = 0;
2114 2115

    if (vm->newDef) {
2116
        virDomainDefFree(vm->def);
2117
        vm->def = vm->newDef;
2118
        vm->def->id = -1;
2119 2120 2121 2122 2123
        vm->newDef = NULL;
    }
}


2124
static void
2125
qemudDispatchVMEvent(int watch, int fd, int events, void *opaque) {
2126
    struct qemud_driver *driver = opaque;
2127
    virDomainObjPtr vm = NULL;
2128
    virDomainEventPtr event = NULL;
2129
    unsigned int i;
2130
    int quit = 0, failed = 0;
2131

2132
    qemuDriverLock(driver);
2133
    for (i = 0 ; i < driver->domains.count ; i++) {
2134 2135 2136
        virDomainObjPtr tmpvm = driver->domains.objs[i];
        virDomainObjLock(tmpvm);
        if (virDomainIsActive(tmpvm) &&
2137
            tmpvm->monitorWatch == watch) {
2138
            vm = tmpvm;
2139
            break;
2140
        }
2141
        virDomainObjUnlock(tmpvm);
2142 2143 2144
    }

    if (!vm)
2145
        goto cleanup;
2146

2147
    if (vm->monitor != fd) {
2148 2149
        failed = 1;
    } else {
2150
        if (events & (VIR_EVENT_HANDLE_HANGUP | VIR_EVENT_HANDLE_ERROR))
2151
            quit = 1;
2152
        else {
2153 2154
            VIR_ERROR(_("unhandled fd event %d for %s"),
                      events, vm->def->name);
2155
            failed = 1;
2156
        }
2157 2158
    }

2159
    if (failed || quit) {
2160 2161 2162 2163 2164
        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_STOPPED,
                                         quit ?
                                         VIR_DOMAIN_EVENT_STOPPED_SHUTDOWN :
                                         VIR_DOMAIN_EVENT_STOPPED_FAILED);
2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175
        qemudShutdownVMDaemon(NULL, driver, vm);
        if (!vm->persistent) {
            virDomainRemoveInactive(&driver->domains,
                                    vm);
            vm = NULL;
        }
    }

cleanup:
    if (vm)
        virDomainObjUnlock(vm);
2176 2177
    if (event)
        qemuDomainEventQueue(driver, event);
2178
    qemuDriverUnlock(driver);
2179 2180
}

2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202

/* Throw away any data available on the monitor
 * This is done before executing a command, in order
 * to allow re-synchronization if something went badly
 * wrong in the past. it also deals with problem of
 * QEMU *sometimes* re-printing its initial greeting
 * when we reconnect to the monitor after restarts.
 */
static void
qemuMonitorDiscardPendingData(virDomainObjPtr vm) {
    char buf[1024];
    int ret = 0;

    /* Monitor is non-blocking, so just loop till we
     * get -1 or 0. Don't bother with detecting
     * errors, since we'll deal with that better later */
    do {
        ret = read(vm->monitor, buf, sizeof (buf)-1);
    } while (ret > 0);
}


2203
static int
2204 2205 2206 2207 2208
qemudMonitorCommandExtra(const virDomainObjPtr vm,
                         const char *cmd,
                         const char *extra,
                         const char *extraPrompt,
                         char **reply) {
D
Daniel P. Berrange 已提交
2209 2210
    int size = 0;
    char *buf = NULL;
2211
    size_t cmdlen = strlen(cmd);
2212
    size_t extralen = extra ? strlen(extra) : 0;
D
Daniel P. Berrange 已提交
2213

2214 2215
    qemuMonitorDiscardPendingData(vm);

2216 2217 2218
    if (safewrite(vm->monitor, cmd, cmdlen) != cmdlen)
        return -1;
    if (safewrite(vm->monitor, "\r", 1) != 1)
D
Daniel P. Berrange 已提交
2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230
        return -1;

    *reply = NULL;

    for (;;) {
        struct pollfd fd = { vm->monitor, POLLIN | POLLERR | POLLHUP, 0 };
        char *tmp;

        /* Read all the data QEMU has sent thus far */
        for (;;) {
            char data[1024];
            int got = read(vm->monitor, data, sizeof(data));
D
Daniel P. Berrange 已提交
2231

2232 2233
            if (got == 0)
                goto error;
D
Daniel P. Berrange 已提交
2234 2235 2236 2237 2238
            if (got < 0) {
                if (errno == EINTR)
                    continue;
                if (errno == EAGAIN)
                    break;
2239
                goto error;
2240
            }
2241
            if (VIR_REALLOC_N(buf, size+got+1) < 0)
2242 2243
                goto error;

D
Daniel P. Berrange 已提交
2244 2245 2246 2247
            memmove(buf+size, data, got);
            buf[size+got] = '\0';
            size += got;
        }
2248

D
Daniel P. Berrange 已提交
2249
        /* Look for QEMU prompt to indicate completion */
2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274
        if (buf) {
            if (extra) {
                if (strstr(buf, extraPrompt) != NULL) {
                    if (safewrite(vm->monitor, extra, extralen) != extralen)
                        return -1;
                    if (safewrite(vm->monitor, "\r", 1) != 1)
                        return -1;
                    extra = NULL;
                }
            } else if ((tmp = strstr(buf, QEMU_CMD_PROMPT)) != NULL) {
                char *commptr = NULL, *nlptr = NULL;
                /* Preserve the newline */
                tmp[1] = '\0';

                /* The monitor doesn't dump clean output after we have written to
                 * it. Every character we write dumps a bunch of useless stuff,
                 * so the result looks like "cXcoXcomXcommXcommaXcommanXcommand"
                 * Try to throw away everything before the first full command
                 * occurence, and inbetween the command and the newline starting
                 * the response
                 */
                if ((commptr = strstr(buf, cmd)))
                    memmove(buf, commptr, strlen(commptr)+1);
                if ((nlptr = strchr(buf, '\n')))
                    memmove(buf+strlen(cmd), nlptr, strlen(nlptr)+1);
C
Cole Robinson 已提交
2275

2276 2277
                break;
            }
D
Daniel P. Berrange 已提交
2278 2279 2280 2281 2282 2283
        }
    pollagain:
        /* Need to wait for more data */
        if (poll(&fd, 1, -1) < 0) {
            if (errno == EINTR)
                goto pollagain;
2284
            goto error;
D
Daniel P. Berrange 已提交
2285 2286 2287 2288
        }
    }
    *reply = buf;
    return 0;
2289 2290

 error:
2291
    VIR_FREE(buf);
2292
    return -1;
D
Daniel P. Berrange 已提交
2293 2294
}

2295 2296 2297 2298 2299 2300 2301 2302
static int
qemudMonitorCommand(const virDomainObjPtr vm,
                    const char *cmd,
                    char **reply) {
    return qemudMonitorCommandExtra(vm, cmd, NULL, NULL, reply);
}


2303

2304
static virDrvOpenStatus qemudOpen(virConnectPtr conn,
2305
                                  virConnectAuthPtr auth ATTRIBUTE_UNUSED,
2306
                                  int flags ATTRIBUTE_UNUSED) {
2307
    if (conn->uri == NULL) {
2308 2309 2310
        if (qemu_driver == NULL)
            return VIR_DRV_OPEN_DECLINED;

2311
        conn->uri = xmlParseURI(qemu_driver->privileged ?
2312 2313
                                "qemu:///system" :
                                "qemu:///session");
2314
        if (!conn->uri) {
2315
            virReportOOMError(conn);
2316 2317
            return VIR_DRV_OPEN_ERROR;
        }
2318 2319 2320 2321 2322 2323 2324 2325 2326 2327
    } else {
        /* If URI isn't 'qemu' its definitely not for us */
        if (conn->uri->scheme == NULL ||
            STRNEQ(conn->uri->scheme, "qemu"))
            return VIR_DRV_OPEN_DECLINED;

        /* Allow remote driver to deal with URIs with hostname server */
        if (conn->uri->server != NULL)
            return VIR_DRV_OPEN_DECLINED;

2328 2329 2330 2331 2332 2333
        if (qemu_driver == NULL) {
            qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR, "%s",
                             _("qemu state driver is not active"));
            return VIR_DRV_OPEN_ERROR;
        }

2334
        if (qemu_driver->privileged) {
2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349
            if (STRNEQ (conn->uri->path, "/system") &&
                STRNEQ (conn->uri->path, "/session")) {
                qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                                 _("unexpected QEMU URI path '%s', try qemu:///system"),
                                 conn->uri->path);
                return VIR_DRV_OPEN_ERROR;
            }
        } else {
            if (STRNEQ (conn->uri->path, "/session")) {
                qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                                 _("unexpected QEMU URI path '%s', try qemu:///session"),
                                 conn->uri->path);
                return VIR_DRV_OPEN_ERROR;
            }
        }
2350 2351 2352 2353 2354 2355 2356
    }
    conn->privateData = qemu_driver;

    return VIR_DRV_OPEN_SUCCESS;
}

static int qemudClose(virConnectPtr conn) {
2357
    struct qemud_driver *driver = conn->privateData;
2358 2359

    /* Get rid of callbacks registered for this conn */
2360
    qemuDriverLock(driver);
2361
    virDomainEventCallbackListRemoveConn(conn, driver->domainEventCallbacks);
2362
    qemuDriverUnlock(driver);
2363 2364 2365 2366 2367 2368

    conn->privateData = NULL;

    return 0;
}

D
Daniel Veillard 已提交
2369 2370 2371 2372 2373 2374 2375 2376 2377 2378
/* Which features are supported by this driver? */
static int
qemudSupportsFeature (virConnectPtr conn ATTRIBUTE_UNUSED, int feature)
{
    switch (feature) {
    case VIR_DRV_FEATURE_MIGRATION_V2: return 1;
    default: return 0;
    }
}

2379
static const char *qemudGetType(virConnectPtr conn ATTRIBUTE_UNUSED) {
2380
    return "QEMU";
2381 2382
}

2383 2384 2385 2386 2387

static int kvmGetMaxVCPUs(void) {
    int maxvcpus = 1;

    int r, fd;
2388

2389 2390
    fd = open(KVM_DEVICE, O_RDONLY);
    if (fd < 0) {
2391 2392
        virReportSystemError(NULL, errno, _("Unable to open %s"), KVM_DEVICE);
        return -1;
2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403
    }

    r = ioctl(fd, KVM_CHECK_EXTENSION, KVM_CAP_NR_VCPUS);
    if (r > 0)
        maxvcpus = r;

    close(fd);
    return maxvcpus;
}


2404
static int qemudGetMaxVCPUs(virConnectPtr conn, const char *type) {
2405 2406 2407
    if (!type)
        return 16;

2408
    if (STRCASEEQ(type, "qemu"))
2409 2410
        return 16;

2411
    if (STRCASEEQ(type, "kvm"))
2412
        return kvmGetMaxVCPUs();
2413

2414
    if (STRCASEEQ(type, "kqemu"))
2415
        return 1;
2416 2417 2418

    qemudReportError(conn, NULL, NULL, VIR_ERR_INVALID_ARG,
                     _("unknown type '%s'"), type);
2419 2420 2421
    return -1;
}

2422

2423
static char *qemudGetCapabilities(virConnectPtr conn) {
2424
    struct qemud_driver *driver = conn->privateData;
2425
    virCapsPtr caps;
2426
    char *xml = NULL;
2427

2428
    qemuDriverLock(driver);
2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440
    if ((caps = qemudCapsInit()) == NULL) {
        virReportOOMError(conn);
        goto cleanup;
    }

    if (qemu_driver->securityDriver &&
        qemudSecurityCapsInit(qemu_driver->securityDriver, caps) < 0) {
        virCapabilitiesFree(caps);
        virReportOOMError(conn);
        goto cleanup;
    }

2441
    virCapabilitiesFree(qemu_driver->caps);
2442 2443 2444
    qemu_driver->caps = caps;

    if ((xml = virCapabilitiesFormatXML(driver->caps)) == NULL)
2445
        virReportOOMError(conn);
2446 2447

cleanup:
2448
    qemuDriverUnlock(driver);
2449

2450
    return xml;
2451 2452 2453
}


D
Daniel P. Berrange 已提交
2454 2455 2456
static int qemudGetProcessInfo(unsigned long long *cpuTime, int pid) {
    char proc[PATH_MAX];
    FILE *pidinfo;
2457
    unsigned long long usertime, systime;
D
Daniel P. Berrange 已提交
2458 2459 2460 2461 2462 2463

    if (snprintf(proc, sizeof(proc), "/proc/%d/stat", pid) >= (int)sizeof(proc)) {
        return -1;
    }

    if (!(pidinfo = fopen(proc, "r"))) {
2464
        /*printf("cannot read pid info");*/
D
Daniel P. Berrange 已提交
2465 2466 2467 2468 2469
        /* VM probably shut down, so fake 0 */
        *cpuTime = 0;
        return 0;
    }

2470
    if (fscanf(pidinfo, "%*d %*s %*c %*d %*d %*d %*d %*d %*u %*u %*u %*u %*u %llu %llu", &usertime, &systime) != 2) {
2471
        qemudDebug("not enough arg");
D
Daniel P. Berrange 已提交
2472 2473 2474 2475 2476 2477 2478 2479
        return -1;
    }

    /* We got jiffies
     * We want nanoseconds
     * _SC_CLK_TCK is jiffies per second
     * So calulate thus....
     */
2480
    *cpuTime = 1000ull * 1000ull * 1000ull * (usertime + systime) / (unsigned long long)sysconf(_SC_CLK_TCK);
D
Daniel P. Berrange 已提交
2481

2482
    qemudDebug("Got %llu %llu %llu", usertime, systime, *cpuTime);
D
Daniel P. Berrange 已提交
2483 2484 2485 2486 2487 2488 2489

    fclose(pidinfo);

    return 0;
}


2490
static virDomainPtr qemudDomainLookupByID(virConnectPtr conn,
2491
                                          int id) {
2492 2493 2494 2495
    struct qemud_driver *driver = conn->privateData;
    virDomainObjPtr vm;
    virDomainPtr dom = NULL;

2496
    qemuDriverLock(driver);
2497
    vm  = virDomainFindByID(&driver->domains, id);
2498
    qemuDriverUnlock(driver);
2499 2500

    if (!vm) {
2501 2502
        qemudReportError(conn, NULL, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching id %d"), id);
2503
        goto cleanup;
2504 2505
    }

2506
    dom = virGetDomain(conn, vm->def->name, vm->def->uuid);
2507
    if (dom) dom->id = vm->def->id;
2508 2509

cleanup:
2510 2511
    if (vm)
        virDomainObjUnlock(vm);
2512 2513
    return dom;
}
2514

2515
static virDomainPtr qemudDomainLookupByUUID(virConnectPtr conn,
2516
                                            const unsigned char *uuid) {
2517 2518 2519
    struct qemud_driver *driver = conn->privateData;
    virDomainObjPtr vm;
    virDomainPtr dom = NULL;
2520

2521
    qemuDriverLock(driver);
2522
    vm = virDomainFindByUUID(&driver->domains, uuid);
2523 2524
    qemuDriverUnlock(driver);

2525
    if (!vm) {
2526 2527 2528
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(uuid, uuidstr);
        qemudReportError(conn, NULL, NULL, VIR_ERR_NO_DOMAIN,
2529
                         _("no domain with matching uuid '%s'"), uuidstr);
2530
        goto cleanup;
2531 2532
    }

2533
    dom = virGetDomain(conn, vm->def->name, vm->def->uuid);
2534
    if (dom) dom->id = vm->def->id;
2535 2536

cleanup:
2537 2538
    if (vm)
        virDomainObjUnlock(vm);
2539 2540
    return dom;
}
2541

2542
static virDomainPtr qemudDomainLookupByName(virConnectPtr conn,
2543
                                            const char *name) {
2544 2545 2546
    struct qemud_driver *driver = conn->privateData;
    virDomainObjPtr vm;
    virDomainPtr dom = NULL;
2547

2548
    qemuDriverLock(driver);
2549
    vm = virDomainFindByName(&driver->domains, name);
2550 2551
    qemuDriverUnlock(driver);

2552
    if (!vm) {
2553 2554
        qemudReportError(conn, NULL, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching name '%s'"), name);
2555
        goto cleanup;
2556 2557
    }

2558
    dom = virGetDomain(conn, vm->def->name, vm->def->uuid);
2559
    if (dom) dom->id = vm->def->id;
2560 2561

cleanup:
2562 2563
    if (vm)
        virDomainObjUnlock(vm);
2564 2565 2566
    return dom;
}

2567
static int qemudGetVersion(virConnectPtr conn, unsigned long *version) {
2568 2569 2570
    struct qemud_driver *driver = conn->privateData;
    int ret = -1;

2571
    qemuDriverLock(driver);
2572
    if (qemudExtractVersion(conn, driver) < 0)
2573
        goto cleanup;
2574

2575
    *version = qemu_driver->qemuVersion;
2576 2577 2578
    ret = 0;

cleanup:
2579
    qemuDriverUnlock(driver);
2580
    return ret;
D
Daniel P. Berrange 已提交
2581 2582
}

2583 2584 2585
static char *
qemudGetHostname (virConnectPtr conn)
{
2586
    char *result;
2587

2588 2589
    result = virGetHostname();
    if (result == NULL) {
2590 2591
        virReportSystemError (conn, errno,
                              "%s", _("failed to determine host name"));
2592 2593 2594
        return NULL;
    }
    /* Caller frees this string. */
2595
    return result;
2596 2597
}

2598
static int qemudListDomains(virConnectPtr conn, int *ids, int nids) {
2599
    struct qemud_driver *driver = conn->privateData;
2600 2601
    int got = 0, i;

2602 2603 2604
    qemuDriverLock(driver);
    for (i = 0 ; i < driver->domains.count && got < nids ; i++) {
        virDomainObjLock(driver->domains.objs[i]);
2605 2606
        if (virDomainIsActive(driver->domains.objs[i]))
            ids[got++] = driver->domains.objs[i]->def->id;
2607 2608 2609
        virDomainObjUnlock(driver->domains.objs[i]);
    }
    qemuDriverUnlock(driver);
2610

D
Daniel P. Berrange 已提交
2611 2612
    return got;
}
2613

2614
static int qemudNumDomains(virConnectPtr conn) {
2615
    struct qemud_driver *driver = conn->privateData;
2616 2617
    int n = 0, i;

2618 2619 2620
    qemuDriverLock(driver);
    for (i = 0 ; i < driver->domains.count ; i++) {
        virDomainObjLock(driver->domains.objs[i]);
2621
        if (virDomainIsActive(driver->domains.objs[i]))
2622
            n++;
2623 2624 2625
        virDomainObjUnlock(driver->domains.objs[i]);
    }
    qemuDriverUnlock(driver);
2626

2627
    return n;
D
Daniel P. Berrange 已提交
2628
}
2629

2630
static virDomainPtr qemudDomainCreate(virConnectPtr conn, const char *xml,
2631
                                      unsigned int flags ATTRIBUTE_UNUSED) {
2632
    struct qemud_driver *driver = conn->privateData;
2633
    virDomainDefPtr def;
2634
    virDomainObjPtr vm = NULL;
2635
    virDomainPtr dom = NULL;
2636
    virDomainEventPtr event = NULL;
D
Daniel P. Berrange 已提交
2637

2638
    qemuDriverLock(driver);
2639 2640
    if (!(def = virDomainDefParseString(conn, driver->caps, xml,
                                        VIR_DOMAIN_XML_INACTIVE)))
2641
        goto cleanup;
2642

2643 2644 2645
    if (virSecurityDriverVerify(conn, def) < 0)
        goto cleanup;

2646
    /* See if a VM with matching UUID already exists */
2647
    vm = virDomainFindByUUID(&driver->domains, def->uuid);
2648
    if (vm) {
2649 2650 2651 2652 2653 2654 2655 2656 2657
        /* UUID matches, but if names don't match, refuse it */
        if (STRNEQ(vm->def->name, def->name)) {
            char uuidstr[VIR_UUID_STRING_BUFLEN];
            virUUIDFormat(vm->def->uuid, uuidstr);
            qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                             _("domain '%s' is already defined with uuid %s"),
                             vm->def->name, uuidstr);
            goto cleanup;
        }
2658

2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676
        /* UUID & name match, but if VM is already active, refuse it */
        if (virDomainIsActive(vm)) {
            qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                             _("domain is already active as '%s'"), vm->def->name);
            goto cleanup;
        }
        virDomainObjUnlock(vm);
    } else {
        /* UUID does not match, but if a name matches, refuse it */
        vm = virDomainFindByName(&driver->domains, def->name);
        if (vm) {
            char uuidstr[VIR_UUID_STRING_BUFLEN];
            virUUIDFormat(vm->def->uuid, uuidstr);
            qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                             _("domain '%s' is already defined with uuid %s"),
                             def->name, uuidstr);
            goto cleanup;
        }
2677
    }
2678

2679 2680
    if (!(vm = virDomainAssignDef(conn,
                                  &driver->domains,
2681 2682 2683 2684
                                  def)))
        goto cleanup;

    def = NULL;
D
Daniel P. Berrange 已提交
2685

2686
    if (qemudStartVMDaemon(conn, driver, vm, NULL, -1) < 0) {
2687 2688
        virDomainRemoveInactive(&driver->domains,
                                vm);
2689
        vm = NULL;
2690
        goto cleanup;
D
Daniel P. Berrange 已提交
2691
    }
2692 2693 2694 2695

    event = virDomainEventNewFromObj(vm,
                                     VIR_DOMAIN_EVENT_STARTED,
                                     VIR_DOMAIN_EVENT_STARTED_BOOTED);
D
Daniel P. Berrange 已提交
2696

2697
    dom = virGetDomain(conn, vm->def->name, vm->def->uuid);
2698
    if (dom) dom->id = vm->def->id;
2699 2700 2701

cleanup:
    virDomainDefFree(def);
2702 2703
    if (vm)
        virDomainObjUnlock(vm);
2704 2705
    if (event)
        qemuDomainEventQueue(driver, event);
2706
    qemuDriverUnlock(driver);
2707
    return dom;
D
Daniel P. Berrange 已提交
2708 2709 2710
}


2711
static int qemudDomainSuspend(virDomainPtr dom) {
2712
    struct qemud_driver *driver = dom->conn->privateData;
D
Daniel P. Berrange 已提交
2713
    char *info;
2714 2715
    virDomainObjPtr vm;
    int ret = -1;
2716
    virDomainEventPtr event = NULL;
2717

2718
    qemuDriverLock(driver);
2719
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
2720

D
Daniel P. Berrange 已提交
2721
    if (!vm) {
2722 2723 2724 2725
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
2726
        goto cleanup;
D
Daniel P. Berrange 已提交
2727
    }
2728
    if (!virDomainIsActive(vm)) {
2729
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
2730
                         "%s", _("domain is not running"));
2731
        goto cleanup;
D
Daniel P. Berrange 已提交
2732
    }
2733
    if (vm->state != VIR_DOMAIN_PAUSED) {
2734
        if (qemudMonitorCommand(vm, "stop", &info) < 0) {
2735 2736 2737 2738 2739 2740
            qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                             "%s", _("suspend operation failed"));
            goto cleanup;
        }
        vm->state = VIR_DOMAIN_PAUSED;
        qemudDebug("Reply %s", info);
2741 2742 2743
        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_SUSPENDED,
                                         VIR_DOMAIN_EVENT_SUSPENDED_PAUSED);
2744
        VIR_FREE(info);
D
Daniel P. Berrange 已提交
2745
    }
2746
    if (virDomainSaveStatus(dom->conn, driver->stateDir, vm) < 0)
2747
        goto cleanup;
2748 2749 2750
    ret = 0;

cleanup:
2751 2752
    if (vm)
        virDomainObjUnlock(vm);
2753

2754
    if (event)
2755
        qemuDomainEventQueue(driver, event);
2756
    qemuDriverUnlock(driver);
2757
    return ret;
D
Daniel P. Berrange 已提交
2758 2759 2760
}


2761
static int qemudDomainResume(virDomainPtr dom) {
2762
    struct qemud_driver *driver = dom->conn->privateData;
D
Daniel P. Berrange 已提交
2763
    char *info;
2764 2765
    virDomainObjPtr vm;
    int ret = -1;
2766
    virDomainEventPtr event = NULL;
2767

2768
    qemuDriverLock(driver);
2769
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
2770

D
Daniel P. Berrange 已提交
2771
    if (!vm) {
2772 2773 2774 2775
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
2776
        goto cleanup;
D
Daniel P. Berrange 已提交
2777
    }
2778
    if (!virDomainIsActive(vm)) {
2779
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
2780
                         "%s", _("domain is not running"));
2781
        goto cleanup;
D
Daniel P. Berrange 已提交
2782
    }
2783
    if (vm->state == VIR_DOMAIN_PAUSED) {
2784
        if (qemudMonitorCommand(vm, "cont", &info) < 0) {
2785 2786 2787 2788 2789 2790
            qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                             "%s", _("resume operation failed"));
            goto cleanup;
        }
        vm->state = VIR_DOMAIN_RUNNING;
        qemudDebug("Reply %s", info);
2791 2792 2793
        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_RESUMED,
                                         VIR_DOMAIN_EVENT_RESUMED_UNPAUSED);
2794
        VIR_FREE(info);
D
Daniel P. Berrange 已提交
2795
    }
2796
    if (virDomainSaveStatus(dom->conn, driver->stateDir, vm) < 0)
2797
        goto cleanup;
2798 2799 2800
    ret = 0;

cleanup:
2801 2802
    if (vm)
        virDomainObjUnlock(vm);
2803
    if (event)
2804
        qemuDomainEventQueue(driver, event);
2805
    qemuDriverUnlock(driver);
2806
    return ret;
D
Daniel P. Berrange 已提交
2807 2808 2809
}


2810
static int qemudDomainShutdown(virDomainPtr dom) {
2811 2812
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
2813
    char* info;
2814
    int ret = -1;
2815

2816
    qemuDriverLock(driver);
2817
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
2818 2819
    qemuDriverUnlock(driver);

2820
    if (!vm) {
2821 2822 2823 2824
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
2825
        goto cleanup;
2826 2827
    }

2828
    if (qemudMonitorCommand(vm, "system_powerdown", &info) < 0) {
2829
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
2830
                         "%s", _("shutdown operation failed"));
2831
        goto cleanup;
2832
    }
D
Daniel Veillard 已提交
2833
    VIR_FREE(info);
2834 2835 2836
    ret = 0;

cleanup:
2837 2838
    if (vm)
        virDomainObjUnlock(vm);
2839
    return ret;
2840 2841 2842
}


2843
static int qemudDomainDestroy(virDomainPtr dom) {
2844 2845 2846
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    int ret = -1;
2847
    virDomainEventPtr event = NULL;
2848

2849
    qemuDriverLock(driver);
2850
    vm  = virDomainFindByUUID(&driver->domains, dom->uuid);
D
Daniel P. Berrange 已提交
2851
    if (!vm) {
2852 2853 2854 2855
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
2856
        goto cleanup;
D
Daniel P. Berrange 已提交
2857
    }
2858 2859 2860 2861 2862
    if (!virDomainIsActive(vm)) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                         "%s", _("domain is not running"));
        goto cleanup;
    }
2863

2864
    qemudShutdownVMDaemon(dom->conn, driver, vm);
2865 2866 2867
    event = virDomainEventNewFromObj(vm,
                                     VIR_DOMAIN_EVENT_STOPPED,
                                     VIR_DOMAIN_EVENT_STOPPED_DESTROYED);
2868
    if (!vm->persistent) {
2869 2870
        virDomainRemoveInactive(&driver->domains,
                                vm);
2871 2872
        vm = NULL;
    }
2873 2874 2875
    ret = 0;

cleanup:
2876 2877
    if (vm)
        virDomainObjUnlock(vm);
2878 2879
    if (event)
        qemuDomainEventQueue(driver, event);
2880
    qemuDriverUnlock(driver);
2881
    return ret;
D
Daniel P. Berrange 已提交
2882 2883 2884
}


2885
static char *qemudDomainGetOSType(virDomainPtr dom) {
2886 2887 2888
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    char *type = NULL;
2889

2890
    qemuDriverLock(driver);
2891
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
2892
    qemuDriverUnlock(driver);
2893
    if (!vm) {
2894 2895 2896 2897
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
2898
        goto cleanup;
2899 2900
    }

2901
    if (!(type = strdup(vm->def->os.type)))
2902
        virReportOOMError(dom->conn);
2903 2904

cleanup:
2905 2906
    if (vm)
        virDomainObjUnlock(vm);
2907 2908 2909
    return type;
}

2910 2911
/* Returns max memory in kb, 0 if error */
static unsigned long qemudDomainGetMaxMemory(virDomainPtr dom) {
2912 2913 2914
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    unsigned long ret = 0;
2915

2916
    qemuDriverLock(driver);
2917
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
2918 2919
    qemuDriverUnlock(driver);

2920
    if (!vm) {
2921 2922
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
2923
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
2924
                         _("no domain with matching uuid '%s'"), uuidstr);
2925
        goto cleanup;
2926 2927
    }

2928 2929 2930
    ret = vm->def->maxmem;

cleanup:
2931 2932
    if (vm)
        virDomainObjUnlock(vm);
2933
    return ret;
2934 2935 2936
}

static int qemudDomainSetMaxMemory(virDomainPtr dom, unsigned long newmax) {
2937 2938 2939
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    int ret = -1;
2940

2941
    qemuDriverLock(driver);
2942
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
2943 2944
    qemuDriverUnlock(driver);

2945
    if (!vm) {
2946 2947
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
2948
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
2949
                         _("no domain with matching uuid '%s'"), uuidstr);
2950
        goto cleanup;
2951 2952 2953 2954
    }

    if (newmax < vm->def->memory) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INVALID_ARG,
2955
                         "%s", _("cannot set max memory lower than current memory"));
2956
        goto cleanup;;
2957 2958 2959
    }

    vm->def->maxmem = newmax;
2960 2961 2962
    ret = 0;

cleanup:
2963 2964
    if (vm)
        virDomainObjUnlock(vm);
2965
    return ret;
2966 2967
}

2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979 2980 2981 2982

/* The reply from QEMU contains 'ballon: actual=421' where value is in MB */
#define BALLOON_PREFIX "balloon: actual="

/*
 * Returns: 0 if balloon not supported, +1 if balloon query worked
 * or -1 on failure
 */
static int qemudDomainGetMemoryBalloon(virConnectPtr conn,
                                       virDomainObjPtr vm,
                                       unsigned long *currmem) {
    char *reply = NULL;
    int ret = -1;
    char *offset;

2983 2984 2985
    if (!virDomainIsActive(vm))
        return 0;

2986 2987 2988 2989 2990 2991
    if (qemudMonitorCommand(vm, "info balloon", &reply) < 0) {
        qemudReportError(conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                         "%s", _("could not query memory balloon allocation"));
        goto cleanup;
    }

2992
    DEBUG ("%s: balloon reply: '%s'", vm->def->name, reply);
2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045
    if ((offset = strstr(reply, BALLOON_PREFIX)) != NULL) {
        unsigned int memMB;
        char *end;
        offset += strlen(BALLOON_PREFIX);
        if (virStrToLong_ui(offset, &end, 10, &memMB) < 0) {
            qemudReportError(conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                             "%s", _("could not parse memory balloon allocation"));
            goto cleanup;
        }
        *currmem = memMB * 1024;
        ret = 1;
    } else {
        /* We don't raise an error here, since its to be expected that
         * many QEMU's don't support ballooning
         */
        ret = 0;
    }

cleanup:
    VIR_FREE(reply);
    return ret;
}

/*
 * Returns: 0 if balloon not supported, +1 if balloon query worked
 * or -1 on failure
 */
static int qemudDomainSetMemoryBalloon(virConnectPtr conn,
                                       virDomainObjPtr vm,
                                       unsigned long newmem) {
    char *cmd;
    char *reply = NULL;
    int ret = -1;

    /*
     * 'newmem' is in KB, QEMU monitor works in MB, and we all wish
     * we just worked in bytes with unsigned long long everywhere.
     */
    if (virAsprintf(&cmd, "balloon %lu", (newmem / 1024)) < 0) {
        virReportOOMError(conn);
        goto cleanup;
    }

    if (qemudMonitorCommand(vm, cmd, &reply) < 0) {
        qemudReportError(conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                         "%s", _("could not balloon memory allocation"));
        VIR_FREE(cmd);
        goto cleanup;
    }
    VIR_FREE(cmd);

    /* If the command failed qemu prints: 'unknown command'
     * No message is printed on success it seems */
3046
    DEBUG ("%s: balloon reply: %s",vm->def->name,  reply);
3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059
    if (strstr(reply, "\nunknown command:")) {
        /* Don't set error - it is expected memory balloon fails on many qemu */
        ret = 0;
    } else {
        ret = 1;
    }

cleanup:
    VIR_FREE(reply);
    return ret;
}


3060
static int qemudDomainSetMemory(virDomainPtr dom, unsigned long newmem) {
3061 3062 3063
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    int ret = -1;
3064

3065
    qemuDriverLock(driver);
3066
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
3067
    qemuDriverUnlock(driver);
3068
    if (!vm) {
3069 3070
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
3071
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
3072
                         _("no domain with matching uuid '%s'"), uuidstr);
3073
        goto cleanup;
3074 3075 3076 3077
    }

    if (newmem > vm->def->maxmem) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INVALID_ARG,
3078
                         "%s", _("cannot set memory higher than max memory"));
3079
        goto cleanup;
3080 3081
    }

3082 3083 3084 3085 3086 3087 3088 3089 3090
    if (virDomainIsActive(vm)) {
        ret = qemudDomainSetMemoryBalloon(dom->conn, vm, newmem);
        if (ret == 0)
            qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                             "%s", _("cannot set memory of an active domain"));
    } else {
        vm->def->memory = newmem;
        ret = 0;
    }
3091 3092

cleanup:
3093 3094
    if (vm)
        virDomainObjUnlock(vm);
3095
    return ret;
3096 3097
}

3098
static int qemudDomainGetInfo(virDomainPtr dom,
3099
                              virDomainInfoPtr info) {
3100 3101 3102
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    int ret = -1;
3103 3104
    int err;
    unsigned long balloon;
3105

3106
    qemuDriverLock(driver);
3107
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
3108
    qemuDriverUnlock(driver);
D
Daniel P. Berrange 已提交
3109
    if (!vm) {
3110 3111 3112 3113
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
3114
        goto cleanup;
D
Daniel P. Berrange 已提交
3115 3116
    }

3117
    info->state = vm->state;
D
Daniel P. Berrange 已提交
3118

3119
    if (!virDomainIsActive(vm)) {
3120
        info->cpuTime = 0;
D
Daniel P. Berrange 已提交
3121
    } else {
3122
        if (qemudGetProcessInfo(&(info->cpuTime), vm->pid) < 0) {
3123
            qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED, ("cannot read cputime for domain"));
3124
            goto cleanup;
D
Daniel P. Berrange 已提交
3125 3126 3127
        }
    }

3128
    info->maxMem = vm->def->maxmem;
3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 3143

    if (virDomainIsActive(vm)) {
        err = qemudDomainGetMemoryBalloon(dom->conn, vm, &balloon);
        if (err < 0)
            goto cleanup;

        if (err == 0)
            /* Balloon not supported, so maxmem is always the allocation */
            info->memory = vm->def->maxmem;
        else
            info->memory = balloon;
    } else {
        info->memory = vm->def->memory;
    }

3144
    info->nrVirtCpu = vm->def->vcpus;
3145 3146 3147
    ret = 0;

cleanup:
3148 3149
    if (vm)
        virDomainObjUnlock(vm);
3150
    return ret;
D
Daniel P. Berrange 已提交
3151 3152 3153
}


D
Daniel P. Berrange 已提交
3154
static char *qemudEscape(const char *in, int shell)
3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175
{
    int len = 0;
    int i, j;
    char *out;

    /* To pass through the QEMU monitor, we need to use escape
       sequences: \r, \n, \", \\

       To pass through both QEMU + the shell, we need to escape
       the single character ' as the five characters '\\''
    */

    for (i = 0; in[i] != '\0'; i++) {
        switch(in[i]) {
        case '\r':
        case '\n':
        case '"':
        case '\\':
            len += 2;
            break;
        case '\'':
D
Daniel P. Berrange 已提交
3176 3177 3178 3179
            if (shell)
                len += 5;
            else
                len += 1;
3180 3181 3182 3183 3184 3185 3186
            break;
        default:
            len += 1;
            break;
        }
    }

3187
    if (VIR_ALLOC_N(out, len + 1) < 0)
3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205
        return NULL;

    for (i = j = 0; in[i] != '\0'; i++) {
        switch(in[i]) {
        case '\r':
            out[j++] = '\\';
            out[j++] = 'r';
            break;
        case '\n':
            out[j++] = '\\';
            out[j++] = 'n';
            break;
        case '"':
        case '\\':
            out[j++] = '\\';
            out[j++] = in[i];
            break;
        case '\'':
D
Daniel P. Berrange 已提交
3206 3207 3208 3209 3210 3211 3212 3213 3214
            if (shell) {
                out[j++] = '\'';
                out[j++] = '\\';
                out[j++] = '\\';
                out[j++] = '\'';
                out[j++] = '\'';
            } else {
                out[j++] = in[i];
            }
3215 3216 3217 3218 3219 3220 3221 3222 3223 3224 3225
            break;
        default:
            out[j++] = in[i];
            break;
        }
    }
    out[j] = '\0';

    return out;
}

3226 3227 3228 3229 3230
static char *qemudEscapeMonitorArg(const char *in)
{
    return qemudEscape(in, 0);
}

D
Daniel P. Berrange 已提交
3231 3232 3233 3234
static char *qemudEscapeShellArg(const char *in)
{
    return qemudEscape(in, 1);
}
3235

3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246
#define QEMUD_SAVE_MAGIC "LibvirtQemudSave"
#define QEMUD_SAVE_VERSION 1

struct qemud_save_header {
    char magic[sizeof(QEMUD_SAVE_MAGIC)-1];
    int version;
    int xml_len;
    int was_running;
    int unused[16];
};

3247
static int qemudDomainSave(virDomainPtr dom,
3248
                           const char *path) {
3249 3250 3251 3252 3253 3254 3255
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    char *command = NULL;
    char *info = NULL;
    int fd = -1;
    char *safe_path = NULL;
    char *xml = NULL;
3256
    struct qemud_save_header header;
3257
    int ret = -1;
3258
    virDomainEventPtr event = NULL;
3259 3260 3261 3262 3263

    memset(&header, 0, sizeof(header));
    memcpy(header.magic, QEMUD_SAVE_MAGIC, sizeof(header.magic));
    header.version = QEMUD_SAVE_VERSION;

3264
    qemuDriverLock(driver);
3265
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
3266

D
Daniel P. Berrange 已提交
3267
    if (!vm) {
3268 3269 3270 3271
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
3272
        goto cleanup;
D
Daniel P. Berrange 已提交
3273
    }
3274

3275
    if (!virDomainIsActive(vm)) {
3276
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
3277
                         "%s", _("domain is not running"));
3278
        goto cleanup;
D
Daniel P. Berrange 已提交
3279
    }
3280 3281 3282 3283

    /* Pause */
    if (vm->state == VIR_DOMAIN_RUNNING) {
        header.was_running = 1;
3284
        if (qemudMonitorCommand(vm, "stop", &info) < 0) {
3285
            qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
3286
                             "%s", _("suspend operation failed"));
3287
            goto cleanup;
3288
        }
3289 3290 3291
        vm->state = VIR_DOMAIN_PAUSED;
        qemudDebug("Reply %s", info);
        VIR_FREE(info);
3292 3293 3294
    }

    /* Get XML for the domain */
3295
    xml = virDomainDefFormat(dom->conn, vm->def, VIR_DOMAIN_XML_SECURE);
3296 3297
    if (!xml) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
3298
                         "%s", _("failed to get domain xml"));
3299
        goto cleanup;
3300 3301 3302 3303 3304 3305
    }
    header.xml_len = strlen(xml) + 1;

    /* Write header to file, followed by XML */
    if ((fd = open(path, O_CREAT|O_TRUNC|O_WRONLY, S_IRUSR|S_IWUSR)) < 0) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
3306
                         _("failed to create '%s'"), path);
3307
        goto cleanup;
3308 3309 3310 3311
    }

    if (safewrite(fd, &header, sizeof(header)) != sizeof(header)) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
3312
                         "%s", _("failed to write save header"));
3313
        goto cleanup;
3314 3315 3316 3317
    }

    if (safewrite(fd, xml, header.xml_len) != header.xml_len) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
3318
                         "%s", _("failed to write xml"));
3319
        goto cleanup;
3320 3321
    }

3322
    if (close(fd) < 0) {
3323 3324 3325
        virReportSystemError(dom->conn, errno,
                             _("unable to save file %s"),
                             path);
3326 3327 3328
        goto cleanup;
    }
    fd = -1;
3329 3330 3331 3332

    /* Migrate to file */
    safe_path = qemudEscapeShellArg(path);
    if (!safe_path) {
3333
        virReportOOMError(dom->conn);
3334
        goto cleanup;
3335
    }
3336
    if (virAsprintf(&command, "migrate \"exec:"
3337
                  "dd of='%s' oflag=append conv=notrunc 2>/dev/null"
3338
                  "\"", safe_path) == -1) {
3339
        virReportOOMError(dom->conn);
3340 3341
        command = NULL;
        goto cleanup;
3342 3343
    }

3344
    if (qemudMonitorCommand(vm, command, &info) < 0) {
3345
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
3346
                         "%s", _("migrate operation failed"));
3347
        goto cleanup;
3348 3349
    }

3350
    DEBUG ("%s: migrate reply: %s", vm->def->name, info);
3351 3352 3353 3354 3355 3356 3357

    /* If the command isn't supported then qemu prints:
     * unknown command: migrate" */
    if (strstr(info, "unknown command:")) {
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                          "%s",
                          _("'migrate' not supported by this qemu"));
3358
        goto cleanup;
3359 3360
    }

3361 3362
    /* Shut it down */
    qemudShutdownVMDaemon(dom->conn, driver, vm);
3363 3364 3365
    event = virDomainEventNewFromObj(vm,
                                     VIR_DOMAIN_EVENT_STOPPED,
                                     VIR_DOMAIN_EVENT_STOPPED_SAVED);
3366
    if (!vm->persistent) {
3367 3368
        virDomainRemoveInactive(&driver->domains,
                                vm);
3369 3370
        vm = NULL;
    }
3371 3372 3373 3374 3375 3376 3377 3378 3379 3380 3381
    ret = 0;

cleanup:
    if (fd != -1)
        close(fd);
    VIR_FREE(xml);
    VIR_FREE(safe_path);
    VIR_FREE(command);
    VIR_FREE(info);
    if (ret != 0)
        unlink(path);
3382 3383
    if (vm)
        virDomainObjUnlock(vm);
3384 3385
    if (event)
        qemuDomainEventQueue(driver, event);
3386
    qemuDriverUnlock(driver);
3387
    return ret;
D
Daniel P. Berrange 已提交
3388 3389 3390
}


P
Paolo Bonzini 已提交
3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447 3448 3449 3450 3451 3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 3468 3469 3470 3471 3472 3473 3474 3475 3476 3477 3478 3479 3480 3481 3482 3483 3484 3485 3486 3487 3488 3489 3490 3491 3492
static int qemudDomainCoreDump(virDomainPtr dom,
                               const char *path,
                               int flags ATTRIBUTE_UNUSED) {
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    char *command = NULL;
    char *info = NULL;
    char *safe_path = NULL;
    int resume = 0, paused = 0;
    int ret = -1;

    qemuDriverLock(driver);
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
    qemuDriverUnlock(driver);

    if (!vm) {
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
        goto cleanup;
    }

    if (!virDomainIsActive(vm)) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
                         "%s", _("domain is not running"));
        goto cleanup;
    }

    /* Migrate will always stop the VM, so once we support live dumping
       the resume condition will stay the same, independent of whether
       the stop command is issued.  */
    resume = (vm->state == VIR_DOMAIN_RUNNING);

    /* Pause domain for non-live dump */
    if (vm->state == VIR_DOMAIN_RUNNING) {
        if (qemudMonitorCommand (vm, "stop", &info) < 0) {
            qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                             "%s", _("suspending before dump failed"));
            goto cleanup;
        }
        DEBUG ("%s: stop reply: %s", vm->def->name, info);
        VIR_FREE(info);
        paused = 1;
    }

    /* Migrate to file */
    safe_path = qemudEscapeShellArg(path);
    if (!safe_path) {
        virReportOOMError(dom->conn);
        goto cleanup;
    }
    if (virAsprintf(&command, "migrate \"exec:"
                  "dd of='%s' 2>/dev/null"
                  "\"", safe_path) == -1) {
        virReportOOMError(dom->conn);
        command = NULL;
        goto cleanup;
    }

    if (qemudMonitorCommand(vm, command, &info) < 0) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                         "%s", _("migrate operation failed"));
        goto cleanup;
    }

    DEBUG ("%s: migrate reply: %s", vm->def->name, info);

    /* If the command isn't supported then qemu prints:
     * unknown command: migrate" */
    if (strstr(info, "unknown command:")) {
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                          "%s",
                          _("'migrate' not supported by this qemu"));
        goto cleanup;
    }

    paused = 1;
    ret = 0;
cleanup:
    VIR_FREE(safe_path);
    VIR_FREE(command);
    VIR_FREE(info);

    /* Since the monitor is always attached to a pty for libvirt, it
       will support synchronous operations so we always get here after
       the migration is complete.  */
    if (resume && paused) {
        if (qemudMonitorCommand(vm, "cont", &info) < 0) {
            qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                             "%s", _("resuming after dump failed"));
            goto cleanup;
        }
        DEBUG ("%s: cont reply: %s", vm->def->name, info);
        VIR_FREE(info);
    }
    if (vm)
        virDomainObjUnlock(vm);
    return ret;
}


3493
static int qemudDomainSetVcpus(virDomainPtr dom, unsigned int nvcpus) {
3494 3495
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
3496
    int max;
3497
    int ret = -1;
3498
    const char *type;
3499

3500
    qemuDriverLock(driver);
3501
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
3502 3503
    qemuDriverUnlock(driver);

3504
    if (!vm) {
3505 3506
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
3507
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
3508
                         _("no domain with matching uuid '%s'"), uuidstr);
3509
        goto cleanup;
3510 3511
    }

3512
    if (virDomainIsActive(vm)) {
3513
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID, "%s",
3514
                         _("cannot change vcpu count of an active domain"));
3515
        goto cleanup;
3516 3517
    }

3518 3519 3520 3521 3522 3523 3524 3525
    if (!(type = virDomainVirtTypeToString(vm->def->virtType))) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("unknown virt type in domain definition '%d'"),
                         vm->def->virtType);
        goto cleanup;
    }

    if ((max = qemudGetMaxVCPUs(dom->conn, type)) < 0) {
3526 3527
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INTERNAL_ERROR, "%s",
                         _("could not determine max vcpus for the domain"));
3528
        goto cleanup;
3529 3530 3531 3532 3533 3534
    }

    if (nvcpus > max) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INVALID_ARG,
                         _("requested vcpus is greater than max allowable"
                           " vcpus for the domain: %d > %d"), nvcpus, max);
3535
        goto cleanup;
3536 3537 3538
    }

    vm->def->vcpus = nvcpus;
3539 3540 3541
    ret = 0;

cleanup:
3542 3543
    if (vm)
        virDomainObjUnlock(vm);
3544
    return ret;
3545 3546
}

3547 3548 3549 3550 3551 3552 3553

#if HAVE_SCHED_GETAFFINITY
static int
qemudDomainPinVcpu(virDomainPtr dom,
                   unsigned int vcpu,
                   unsigned char *cpumap,
                   int maplen) {
3554 3555
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
3556 3557 3558
    cpu_set_t mask;
    int i, maxcpu;
    virNodeInfo nodeinfo;
3559
    int ret = -1;
3560

3561
    qemuDriverLock(driver);
3562
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
3563 3564
    qemuDriverUnlock(driver);

3565 3566 3567 3568 3569 3570 3571 3572
    if (!vm) {
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
        goto cleanup;
    }

3573
    if (!virDomainIsActive(vm)) {
3574
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
3575
                         "%s",_("cannot pin vcpus on an inactive domain"));
3576
        goto cleanup;
3577 3578 3579 3580 3581 3582
    }

    if (vcpu > (vm->nvcpupids-1)) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INVALID_ARG,
                         _("vcpu number out of range %d > %d"),
                         vcpu, vm->nvcpupids);
3583
        goto cleanup;
3584 3585
    }

3586
    if (nodeGetInfo(dom->conn, &nodeinfo) < 0)
3587
        goto cleanup;
3588 3589 3590 3591 3592 3593 3594 3595 3596 3597 3598 3599 3600

    maxcpu = maplen * 8;
    if (maxcpu > nodeinfo.cpus)
        maxcpu = nodeinfo.cpus;

    CPU_ZERO(&mask);
    for (i = 0 ; i < maxcpu ; i++) {
        if ((cpumap[i/8] >> (i % 8)) & 1)
            CPU_SET(i, &mask);
    }

    if (vm->vcpupids != NULL) {
        if (sched_setaffinity(vm->vcpupids[vcpu], sizeof(mask), &mask) < 0) {
3601 3602
            virReportSystemError(dom->conn, errno, "%s",
                                 _("cannot set affinity"));
3603
            goto cleanup;
3604 3605 3606 3607
        }
    } else {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                         "%s", _("cpu affinity is not supported"));
3608
        goto cleanup;
3609
    }
3610
    ret = 0;
3611

3612
cleanup:
3613 3614
    if (vm)
        virDomainObjUnlock(vm);
3615
    return ret;
3616 3617 3618 3619 3620 3621 3622 3623
}

static int
qemudDomainGetVcpus(virDomainPtr dom,
                    virVcpuInfoPtr info,
                    int maxinfo,
                    unsigned char *cpumaps,
                    int maplen) {
3624 3625
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
3626 3627
    virNodeInfo nodeinfo;
    int i, v, maxcpu;
3628
    int ret = -1;
3629

3630
    qemuDriverLock(driver);
3631
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
3632 3633
    qemuDriverUnlock(driver);

3634 3635 3636 3637 3638 3639 3640 3641
    if (!vm) {
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
        goto cleanup;
    }

3642
    if (!virDomainIsActive(vm)) {
3643
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
3644
                         "%s",_("cannot pin vcpus on an inactive domain"));
3645
        goto cleanup;
3646 3647
    }

3648
    if (nodeGetInfo(dom->conn, &nodeinfo) < 0)
3649
        goto cleanup;
3650 3651 3652 3653 3654 3655 3656 3657 3658

    maxcpu = maplen * 8;
    if (maxcpu > nodeinfo.cpus)
        maxcpu = nodeinfo.cpus;

    /* Clamp to actual number of vcpus */
    if (maxinfo > vm->nvcpupids)
        maxinfo = vm->nvcpupids;

3659 3660 3661 3662 3663 3664 3665 3666
    if (maxinfo >= 1) {
        if (info != NULL) {
            memset(info, 0, sizeof(*info) * maxinfo);
            for (i = 0 ; i < maxinfo ; i++) {
                info[i].number = i;
                info[i].state = VIR_VCPU_RUNNING;
                /* XXX cpu time, current pCPU mapping */
            }
3667 3668
        }

3669 3670 3671 3672 3673 3674 3675 3676 3677
        if (cpumaps != NULL) {
            memset(cpumaps, 0, maplen * maxinfo);
            if (vm->vcpupids != NULL) {
                for (v = 0 ; v < maxinfo ; v++) {
                    cpu_set_t mask;
                    unsigned char *cpumap = VIR_GET_CPUMAP(cpumaps, maplen, v);
                    CPU_ZERO(&mask);

                    if (sched_getaffinity(vm->vcpupids[v], sizeof(mask), &mask) < 0) {
3678 3679
                        virReportSystemError(dom->conn, errno, "%s",
                                             _("cannot get affinity"));
3680 3681 3682 3683 3684 3685
                        goto cleanup;
                    }

                    for (i = 0 ; i < maxcpu ; i++)
                        if (CPU_ISSET(i, &mask))
                            VIR_USE_CPU(cpumap, i);
3686
                }
3687 3688 3689 3690
            } else {
                qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                                 "%s", _("cpu affinity is not available"));
                goto cleanup;
3691 3692 3693
            }
        }
    }
3694
    ret = maxinfo;
3695

3696
cleanup:
3697 3698
    if (vm)
        virDomainObjUnlock(vm);
3699
    return ret;
3700 3701 3702 3703
}
#endif /* HAVE_SCHED_GETAFFINITY */


3704
static int qemudDomainGetMaxVcpus(virDomainPtr dom) {
3705 3706
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
3707
    const char *type;
3708
    int ret = -1;
3709

3710
    qemuDriverLock(driver);
3711
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
3712 3713
    qemuDriverUnlock(driver);

3714
    if (!vm) {
3715 3716
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
3717
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
3718
                         _("no domain with matching uuid '%s'"), uuidstr);
3719
        goto cleanup;
3720 3721
    }

3722
    if (!(type = virDomainVirtTypeToString(vm->def->virtType))) {
3723 3724 3725
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("unknown virt type in domain definition '%d'"),
                         vm->def->virtType);
3726
        goto cleanup;
3727 3728
    }

3729
    ret = qemudGetMaxVCPUs(dom->conn, type);
3730

3731
cleanup:
3732 3733
    if (vm)
        virDomainObjUnlock(vm);
3734 3735 3736
    return ret;
}

3737 3738 3739 3740 3741 3742 3743 3744 3745 3746
static int qemudDomainGetSecurityLabel(virDomainPtr dom, virSecurityLabelPtr seclabel)
{
    struct qemud_driver *driver = (struct qemud_driver *)dom->conn->privateData;
    virDomainObjPtr vm;
    const char *type;
    int ret = -1;

    qemuDriverLock(driver);
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);

3747 3748
    memset(seclabel, 0, sizeof(*seclabel));

3749 3750 3751
    if (!vm) {
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
3752
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
3753 3754 3755 3756 3757 3758 3759 3760 3761 3762 3763 3764 3765 3766 3767 3768 3769 3770 3771 3772 3773 3774 3775 3776 3777 3778 3779 3780 3781
                         _("no domain with matching uuid '%s'"), uuidstr);
        goto cleanup;
    }

    if (!(type = virDomainVirtTypeToString(vm->def->virtType))) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("unknown virt type in domain definition '%d'"),
                         vm->def->virtType);
        goto cleanup;
    }

    /*
     * Theoretically, the pid can be replaced during this operation and
     * return the label of a different process.  If atomicity is needed,
     * further validation will be required.
     *
     * Comment from Dan Berrange:
     *
     *   Well the PID as stored in the virDomainObjPtr can't be changed
     *   because you've got a locked object.  The OS level PID could have
     *   exited, though and in extreme circumstances have cycled through all
     *   PIDs back to ours. We could sanity check that our PID still exists
     *   after reading the label, by checking that our FD connecting to the
     *   QEMU monitor hasn't seen SIGHUP/ERR on poll().
     */
    if (virDomainIsActive(vm)) {
        if (driver->securityDriver && driver->securityDriver->domainGetSecurityLabel) {
            if (driver->securityDriver->domainGetSecurityLabel(dom->conn, vm, seclabel) == -1) {
                qemudReportError(dom->conn, dom, NULL, VIR_ERR_INTERNAL_ERROR,
3782
                                 "%s", _("Failed to get security label"));
3783 3784 3785 3786 3787 3788 3789 3790 3791 3792
                goto cleanup;
            }
        }
    }

    ret = 0;

cleanup:
    if (vm)
        virDomainObjUnlock(vm);
3793
    qemuDriverUnlock(driver);
3794 3795 3796
    return ret;
}

3797 3798
static int qemudNodeGetSecurityModel(virConnectPtr conn,
                                     virSecurityModelPtr secmodel)
3799 3800 3801
{
    struct qemud_driver *driver = (struct qemud_driver *)conn->privateData;
    char *p;
3802
    int ret = 0;
3803

3804 3805
    qemuDriverLock(driver);
    if (!driver->securityDriver) {
3806
        memset(secmodel, 0, sizeof (*secmodel));
3807 3808
        goto cleanup;
    }
3809

3810 3811 3812 3813 3814
    p = driver->caps->host.secModel.model;
    if (strlen(p) >= VIR_SECURITY_MODEL_BUFLEN-1) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("security model string exceeds max %d bytes"),
                         VIR_SECURITY_MODEL_BUFLEN-1);
3815 3816
        ret = -1;
        goto cleanup;
3817 3818 3819 3820 3821 3822 3823 3824
    }
    strcpy(secmodel->model, p);

    p = driver->caps->host.secModel.doi;
    if (strlen(p) >= VIR_SECURITY_DOI_BUFLEN-1) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("security DOI string exceeds max %d bytes"),
                         VIR_SECURITY_DOI_BUFLEN-1);
3825 3826
        ret = -1;
        goto cleanup;
3827 3828
    }
    strcpy(secmodel->doi, p);
3829 3830 3831 3832

cleanup:
    qemuDriverUnlock(driver);
    return ret;
3833 3834 3835
}

/* TODO: check seclabel restore */
3836
static int qemudDomainRestore(virConnectPtr conn,
3837 3838 3839
                              const char *path) {
    struct qemud_driver *driver = conn->privateData;
    virDomainDefPtr def = NULL;
3840
    virDomainObjPtr vm = NULL;
3841 3842 3843
    int fd = -1;
    int ret = -1;
    char *xml = NULL;
3844
    struct qemud_save_header header;
3845
    virDomainEventPtr event = NULL;
3846

3847
    qemuDriverLock(driver);
3848 3849 3850
    /* Verify the header and read the XML */
    if ((fd = open(path, O_RDONLY)) < 0) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
3851
                         "%s", _("cannot read domain image"));
3852
        goto cleanup;
3853 3854 3855 3856
    }

    if (saferead(fd, &header, sizeof(header)) != sizeof(header)) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
3857
                         "%s", _("failed to read qemu header"));
3858
        goto cleanup;
3859 3860 3861 3862
    }

    if (memcmp(header.magic, QEMUD_SAVE_MAGIC, sizeof(header.magic)) != 0) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
3863
                         "%s", _("image magic is incorrect"));
3864
        goto cleanup;
3865 3866 3867 3868
    }

    if (header.version > QEMUD_SAVE_VERSION) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
3869
                         _("image version is not supported (%d > %d)"),
3870
                         header.version, QEMUD_SAVE_VERSION);
3871
        goto cleanup;
3872 3873
    }

3874
    if (VIR_ALLOC_N(xml, header.xml_len) < 0) {
3875
        virReportOOMError(conn);
3876
        goto cleanup;
3877 3878 3879 3880
    }

    if (saferead(fd, xml, header.xml_len) != header.xml_len) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
3881
                         "%s", _("failed to read XML"));
3882
        goto cleanup;
3883 3884 3885
    }

    /* Create a domain from this XML */
3886 3887
    if (!(def = virDomainDefParseString(conn, driver->caps, xml,
                                        VIR_DOMAIN_XML_INACTIVE))) {
3888
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
3889
                         "%s", _("failed to parse XML"));
3890
        goto cleanup;
3891 3892
    }

3893
    /* See if a VM with matching UUID already exists */
3894
    vm = virDomainFindByUUID(&driver->domains, def->uuid);
3895
    if (vm) {
3896 3897 3898 3899 3900 3901 3902 3903 3904 3905 3906
        /* UUID matches, but if names don't match, refuse it */
        if (STRNEQ(vm->def->name, def->name)) {
            char uuidstr[VIR_UUID_STRING_BUFLEN];
            virUUIDFormat(vm->def->uuid, uuidstr);
            qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                             _("domain '%s' is already defined with uuid %s"),
                             vm->def->name, uuidstr);
            goto cleanup;
        }

        /* UUID & name match, but if VM is already active, refuse it */
3907
        if (virDomainIsActive(vm)) {
3908
            qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_INVALID,
3909 3910
                             _("domain is already active as '%s'"), vm->def->name);
            goto cleanup;
3911 3912 3913 3914 3915 3916 3917 3918 3919 3920 3921 3922
        }
        virDomainObjUnlock(vm);
    } else {
        /* UUID does not match, but if a name matches, refuse it */
        vm = virDomainFindByName(&driver->domains, def->name);
        if (vm) {
            char uuidstr[VIR_UUID_STRING_BUFLEN];
            virUUIDFormat(vm->def->uuid, uuidstr);
            qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                             _("domain '%s' is already defined with uuid %s"),
                             def->name, uuidstr);
            goto cleanup;
3923
        }
3924 3925
    }

3926 3927 3928
    if (!(vm = virDomainAssignDef(conn,
                                  &driver->domains,
                                  def))) {
3929
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
3930
                         "%s", _("failed to assign new VM"));
3931
        goto cleanup;
3932
    }
3933
    def = NULL;
3934 3935

    /* Set the migration source and start it up. */
3936
    ret = qemudStartVMDaemon(conn, driver, vm, "stdio", fd);
3937
    close(fd);
3938
    fd = -1;
3939
    if (ret < 0) {
3940
        if (!vm->persistent) {
3941 3942
            virDomainRemoveInactive(&driver->domains,
                                    vm);
3943 3944
            vm = NULL;
        }
3945
        goto cleanup;
3946 3947
    }

3948 3949 3950
    event = virDomainEventNewFromObj(vm,
                                     VIR_DOMAIN_EVENT_STARTED,
                                     VIR_DOMAIN_EVENT_STARTED_RESTORED);
3951

3952 3953 3954
    /* If it was running before, resume it now. */
    if (header.was_running) {
        char *info;
3955
        if (qemudMonitorCommand(vm, "cont", &info) < 0) {
3956
            qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
3957
                             "%s", _("failed to resume domain"));
3958
            goto cleanup;
3959
        }
3960
        VIR_FREE(info);
3961 3962
        vm->state = VIR_DOMAIN_RUNNING;
    }
3963
    ret = 0;
3964

3965 3966 3967 3968 3969
cleanup:
    virDomainDefFree(def);
    VIR_FREE(xml);
    if (fd != -1)
        close(fd);
3970 3971
    if (vm)
        virDomainObjUnlock(vm);
3972 3973
    if (event)
        qemuDomainEventQueue(driver, event);
3974
    qemuDriverUnlock(driver);
3975
    return ret;
D
Daniel P. Berrange 已提交
3976 3977 3978
}


3979
static char *qemudDomainDumpXML(virDomainPtr dom,
3980
                                int flags) {
3981 3982 3983
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    char *ret = NULL;
3984 3985
    unsigned long balloon;
    int err;
3986

3987
    qemuDriverLock(driver);
3988
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
3989 3990
    qemuDriverUnlock(driver);

D
Daniel P. Berrange 已提交
3991
    if (!vm) {
3992 3993 3994 3995
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
3996
        goto cleanup;
D
Daniel P. Berrange 已提交
3997 3998
    }

3999 4000 4001 4002 4003 4004 4005
    /* Refresh current memory based on balloon info */
    err = qemudDomainGetMemoryBalloon(dom->conn, vm, &balloon);
    if (err < 0)
        goto cleanup;
    if (err > 0)
        vm->def->memory = balloon;

4006 4007 4008 4009 4010 4011
    ret = virDomainDefFormat(dom->conn,
                             (flags & VIR_DOMAIN_XML_INACTIVE) && vm->newDef ?
                             vm->newDef : vm->def,
                             flags);

cleanup:
4012 4013
    if (vm)
        virDomainObjUnlock(vm);
4014
    return ret;
D
Daniel P. Berrange 已提交
4015 4016 4017
}


4018 4019 4020 4021
static char *qemuDomainXMLFromNative(virConnectPtr conn,
                                     const char *format,
                                     const char *config,
                                     unsigned int flags ATTRIBUTE_UNUSED) {
4022
    struct qemud_driver *driver = conn->privateData;
4023 4024 4025 4026 4027 4028 4029 4030 4031
    virDomainDefPtr def = NULL;
    char *xml = NULL;

    if (STRNEQ(format, QEMU_CONFIG_FORMAT_ARGV)) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_INVALID_ARG,
                         _("unsupported config type %s"), format);
        goto cleanup;
    }

4032
    def = qemuParseCommandLineString(conn, driver->caps, config);
4033 4034 4035 4036 4037 4038 4039 4040 4041 4042
    if (!def)
        goto cleanup;

    xml = virDomainDefFormat(conn, def, VIR_DOMAIN_XML_INACTIVE);

cleanup:
    virDomainDefFree(def);
    return xml;
}

4043 4044 4045 4046 4047 4048
static char *qemuDomainXMLToNative(virConnectPtr conn,
                                   const char *format,
                                   const char *xmlData,
                                   unsigned int flags ATTRIBUTE_UNUSED) {
    struct qemud_driver *driver = conn->privateData;
    virDomainDefPtr def = NULL;
4049
    virDomainChrDef monitor_chr;
4050 4051 4052 4053 4054 4055 4056 4057 4058 4059
    const char *emulator;
    unsigned int qemuCmdFlags;
    struct stat sb;
    const char **retargv = NULL;
    const char **retenv = NULL;
    const char **tmp;
    virBuffer buf = VIR_BUFFER_INITIALIZER;
    char *ret = NULL;
    int i;

4060 4061
    qemuDriverLock(driver);

4062 4063 4064 4065 4066 4067 4068 4069 4070 4071 4072 4073 4074 4075 4076 4077 4078 4079 4080 4081 4082 4083 4084 4085 4086 4087 4088 4089 4090 4091 4092 4093 4094 4095 4096 4097 4098 4099 4100 4101 4102 4103 4104 4105 4106 4107 4108 4109 4110 4111 4112 4113 4114 4115 4116 4117 4118 4119 4120 4121 4122 4123 4124 4125 4126 4127
    if (STRNEQ(format, QEMU_CONFIG_FORMAT_ARGV)) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_INVALID_ARG,
                         _("unsupported config type %s"), format);
        goto cleanup;
    }

    def = virDomainDefParseString(conn, driver->caps, xmlData, 0);
    if (!def)
        goto cleanup;

    /* Since we're just exporting args, we can't do bridge/network
     * setups, since libvirt will normally create TAP devices
     * directly. We convert those configs into generic 'ethernet'
     * config and assume the user has suitable 'ifup-qemu' scripts
     */
    for (i = 0 ; i < def->nnets ; i++) {
        virDomainNetDefPtr net = def->nets[i];
        if (net->type == VIR_DOMAIN_NET_TYPE_NETWORK) {
            VIR_FREE(net->data.network.name);

            memset(net, 0, sizeof *net);

            net->type = VIR_DOMAIN_NET_TYPE_ETHERNET;
            net->data.ethernet.dev = NULL;
            net->data.ethernet.script = NULL;
            net->data.ethernet.ipaddr = NULL;
        } else if (net->type == VIR_DOMAIN_NET_TYPE_BRIDGE) {
            char *brname = net->data.bridge.brname;
            char *script = net->data.bridge.script;
            char *ipaddr = net->data.bridge.ipaddr;

            memset(net, 0, sizeof *net);

            net->type = VIR_DOMAIN_NET_TYPE_ETHERNET;
            net->data.ethernet.dev = brname;
            net->data.ethernet.script = script;
            net->data.ethernet.ipaddr = ipaddr;
        }
    }
    for (i = 0 ; i < def->ngraphics ; i++) {
        if (def->graphics[i]->type == VIR_DOMAIN_GRAPHICS_TYPE_VNC &&
            def->graphics[i]->data.vnc.autoport)
            def->graphics[i]->data.vnc.port = 5900;
    }
    emulator = def->emulator;

    /* Make sure the binary we are about to try exec'ing exists.
     * Technically we could catch the exec() failure, but that's
     * in a sub-process so its hard to feed back a useful error
     */
    if (stat(emulator, &sb) < 0) {
        virReportSystemError(conn, errno,
                             _("Cannot find QEMU binary %s"),
                             emulator);
        goto cleanup;
    }

    if (qemudExtractVersionInfo(emulator,
                                NULL,
                                &qemuCmdFlags) < 0) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("Cannot determine QEMU argv syntax %s"),
                         emulator);
        goto cleanup;
    }

4128 4129
    if (qemuPrepareMonitorChr(conn, driver, &monitor_chr, def->name) < 0)
        goto cleanup;
4130 4131

    if (qemudBuildCommandLine(conn, driver, def,
4132
                              &monitor_chr, qemuCmdFlags,
4133 4134 4135 4136 4137 4138 4139 4140 4141 4142 4143 4144 4145 4146 4147 4148 4149 4150 4151 4152 4153 4154 4155 4156 4157
                              &retargv, &retenv,
                              NULL, NULL, /* Don't want it to create TAP devices */
                              NULL) < 0) {
        goto cleanup;
    }

    tmp = retenv;
    while (*tmp) {
        virBufferAdd(&buf, *tmp, strlen(*tmp));
        virBufferAddLit(&buf, " ");
        tmp++;
    }
    tmp = retargv;
    while (*tmp) {
        virBufferAdd(&buf, *tmp, strlen(*tmp));
        virBufferAddLit(&buf, " ");
        tmp++;
    }

    if (virBufferError(&buf))
        goto cleanup;

    ret = virBufferContentAndReset(&buf);

cleanup:
4158
    qemuDriverUnlock(driver);
4159 4160 4161 4162 4163 4164 4165 4166 4167 4168 4169 4170 4171
    for (tmp = retargv ; tmp && *tmp ; tmp++)
        VIR_FREE(*tmp);
    VIR_FREE(retargv);

    for (tmp = retenv ; tmp && *tmp ; tmp++)
        VIR_FREE(*tmp);
    VIR_FREE(retenv);

    virDomainDefFree(def);
    return ret;
}


4172
static int qemudListDefinedDomains(virConnectPtr conn,
4173
                            char **const names, int nnames) {
4174
    struct qemud_driver *driver = conn->privateData;
4175
    int got = 0, i;
4176

4177
    qemuDriverLock(driver);
4178
    for (i = 0 ; i < driver->domains.count && got < nnames ; i++) {
4179
        virDomainObjLock(driver->domains.objs[i]);
4180 4181
        if (!virDomainIsActive(driver->domains.objs[i])) {
            if (!(names[got++] = strdup(driver->domains.objs[i]->def->name))) {
4182
                virReportOOMError(conn);
4183
                virDomainObjUnlock(driver->domains.objs[i]);
4184 4185
                goto cleanup;
            }
4186
        }
4187
        virDomainObjUnlock(driver->domains.objs[i]);
D
Daniel P. Berrange 已提交
4188
    }
4189

4190
    qemuDriverUnlock(driver);
D
Daniel P. Berrange 已提交
4191
    return got;
4192 4193 4194

 cleanup:
    for (i = 0 ; i < got ; i++)
4195
        VIR_FREE(names[i]);
4196
    qemuDriverUnlock(driver);
4197
    return -1;
D
Daniel P. Berrange 已提交
4198 4199
}

4200
static int qemudNumDefinedDomains(virConnectPtr conn) {
4201
    struct qemud_driver *driver = conn->privateData;
4202 4203
    int n = 0, i;

4204
    qemuDriverLock(driver);
4205 4206
    for (i = 0 ; i < driver->domains.count ; i++)
        if (!virDomainIsActive(driver->domains.objs[i]))
4207
            n++;
4208
    qemuDriverUnlock(driver);
4209

4210
    return n;
D
Daniel P. Berrange 已提交
4211 4212 4213
}


4214
static int qemudDomainStart(virDomainPtr dom) {
4215 4216 4217
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    int ret = -1;
4218
    virDomainEventPtr event = NULL;
4219

4220
    qemuDriverLock(driver);
4221
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
4222

4223
    if (!vm) {
4224 4225 4226 4227
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
4228
        goto cleanup;
4229 4230
    }

4231
    ret = qemudStartVMDaemon(dom->conn, driver, vm, NULL, -1);
4232
    if (ret != -1)
4233 4234 4235
        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_STARTED,
                                         VIR_DOMAIN_EVENT_STARTED_BOOTED);
4236 4237

cleanup:
4238 4239
    if (vm)
        virDomainObjUnlock(vm);
4240
    if (event)
4241
        qemuDomainEventQueue(driver, event);
4242
    qemuDriverUnlock(driver);
4243
    return ret;
D
Daniel P. Berrange 已提交
4244 4245 4246
}


4247
static virDomainPtr qemudDomainDefine(virConnectPtr conn, const char *xml) {
4248
    struct qemud_driver *driver = conn->privateData;
4249
    virDomainDefPtr def;
4250
    virDomainObjPtr vm = NULL;
4251
    virDomainPtr dom = NULL;
4252
    virDomainEventPtr event = NULL;
4253
    int newVM = 1;
4254

4255
    qemuDriverLock(driver);
4256 4257
    if (!(def = virDomainDefParseString(conn, driver->caps, xml,
                                        VIR_DOMAIN_XML_INACTIVE)))
4258
        goto cleanup;
4259

4260 4261 4262
    if (virSecurityDriverVerify(conn, def) < 0)
        goto cleanup;

4263 4264
    /* See if a VM with matching UUID already exists */
    vm = virDomainFindByUUID(&driver->domains, def->uuid);
4265
    if (vm) {
4266 4267 4268 4269 4270 4271 4272 4273 4274 4275 4276
        /* UUID matches, but if names don't match, refuse it */
        if (STRNEQ(vm->def->name, def->name)) {
            char uuidstr[VIR_UUID_STRING_BUFLEN];
            virUUIDFormat(vm->def->uuid, uuidstr);
            qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                             _("domain '%s' is already defined with uuid %s"),
                             vm->def->name, uuidstr);
            goto cleanup;
        }

        /* UUID & name match */
4277
        virDomainObjUnlock(vm);
4278
        newVM = 0;
4279 4280 4281 4282 4283 4284 4285 4286 4287 4288 4289
    } else {
        /* UUID does not match, but if a name matches, refuse it */
        vm = virDomainFindByName(&driver->domains, def->name);
        if (vm) {
            char uuidstr[VIR_UUID_STRING_BUFLEN];
            virUUIDFormat(vm->def->uuid, uuidstr);
            qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                             _("domain '%s' is already defined with uuid %s"),
                             def->name, uuidstr);
            goto cleanup;
        }
4290
    }
4291

4292 4293 4294
    if (!(vm = virDomainAssignDef(conn,
                                  &driver->domains,
                                  def))) {
4295
        goto cleanup;
4296
    }
4297
    def = NULL;
4298
    vm->persistent = 1;
4299

4300 4301
    if (virDomainSaveConfig(conn,
                            driver->configDir,
4302
                            vm->newDef ? vm->newDef : vm->def) < 0) {
4303 4304
        virDomainRemoveInactive(&driver->domains,
                                vm);
4305
        vm = NULL;
4306
        goto cleanup;
4307 4308
    }

4309 4310 4311 4312 4313
    event = virDomainEventNewFromObj(vm,
                                     VIR_DOMAIN_EVENT_DEFINED,
                                     newVM ?
                                     VIR_DOMAIN_EVENT_DEFINED_ADDED :
                                     VIR_DOMAIN_EVENT_DEFINED_UPDATED);
4314

4315
    dom = virGetDomain(conn, vm->def->name, vm->def->uuid);
4316
    if (dom) dom->id = vm->def->id;
4317 4318

cleanup:
4319
    virDomainDefFree(def);
4320 4321
    if (vm)
        virDomainObjUnlock(vm);
4322 4323
    if (event)
        qemuDomainEventQueue(driver, event);
4324
    qemuDriverUnlock(driver);
4325
    return dom;
D
Daniel P. Berrange 已提交
4326 4327
}

4328
static int qemudDomainUndefine(virDomainPtr dom) {
4329 4330
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
4331
    virDomainEventPtr event = NULL;
4332
    int ret = -1;
D
Daniel P. Berrange 已提交
4333

4334
    qemuDriverLock(driver);
4335
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
4336

D
Daniel P. Berrange 已提交
4337
    if (!vm) {
4338 4339 4340 4341
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
4342
        goto cleanup;
D
Daniel P. Berrange 已提交
4343 4344
    }

4345
    if (virDomainIsActive(vm)) {
4346
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
4347
                         "%s", _("cannot delete active domain"));
4348
        goto cleanup;
D
Daniel P. Berrange 已提交
4349 4350
    }

4351 4352 4353
    if (!vm->persistent) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INTERNAL_ERROR,
                         "%s", _("cannot undefine transient domain"));
4354
        goto cleanup;
4355 4356 4357
    }

    if (virDomainDeleteConfig(dom->conn, driver->configDir, driver->autostartDir, vm) < 0)
4358
        goto cleanup;
D
Daniel P. Berrange 已提交
4359

4360 4361 4362
    event = virDomainEventNewFromObj(vm,
                                     VIR_DOMAIN_EVENT_UNDEFINED,
                                     VIR_DOMAIN_EVENT_UNDEFINED_REMOVED);
4363

4364 4365
    virDomainRemoveInactive(&driver->domains,
                            vm);
4366
    vm = NULL;
4367
    ret = 0;
D
Daniel P. Berrange 已提交
4368

4369
cleanup:
4370 4371
    if (vm)
        virDomainObjUnlock(vm);
4372 4373
    if (event)
        qemuDomainEventQueue(driver, event);
4374
    qemuDriverUnlock(driver);
4375
    return ret;
D
Daniel P. Berrange 已提交
4376 4377
}

4378
/* Return the disks name for use in monitor commands */
4379
static char *qemudDiskDeviceName(const virConnectPtr conn,
4380
                                 const virDomainDiskDefPtr disk) {
4381 4382 4383 4384 4385 4386

    int busid, devid;
    int ret;
    char *devname;

    if (virDiskNameToBusDeviceIndex(disk, &busid, &devid) < 0) {
4387
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
4388 4389 4390 4391 4392 4393 4394
                         _("cannot convert disk '%s' to bus/device index"),
                         disk->dst);
        return NULL;
    }

    switch (disk->bus) {
        case VIR_DOMAIN_DISK_BUS_IDE:
4395
            if (disk->device== VIR_DOMAIN_DISK_DEVICE_DISK)
4396
                ret = virAsprintf(&devname, "ide%d-hd%d", busid, devid);
4397
            else
4398
                ret = virAsprintf(&devname, "ide%d-cd%d", busid, devid);
4399 4400
            break;
        case VIR_DOMAIN_DISK_BUS_SCSI:
4401
            if (disk->device == VIR_DOMAIN_DISK_DEVICE_DISK)
4402
                ret = virAsprintf(&devname, "scsi%d-hd%d", busid, devid);
4403
            else
4404
                ret = virAsprintf(&devname, "scsi%d-cd%d", busid, devid);
4405 4406
            break;
        case VIR_DOMAIN_DISK_BUS_FDC:
4407
            ret = virAsprintf(&devname, "floppy%d", devid);
4408 4409
            break;
        case VIR_DOMAIN_DISK_BUS_VIRTIO:
4410
            ret = virAsprintf(&devname, "virtio%d", devid);
4411 4412
            break;
        default:
4413
            qemudReportError(conn, NULL, NULL, VIR_ERR_NO_SUPPORT,
4414 4415 4416 4417 4418 4419
                             _("Unsupported disk name mapping for bus '%s'"),
                             virDomainDiskBusTypeToString(disk->bus));
            return NULL;
    }

    if (ret == -1) {
4420
        virReportOOMError(conn);
4421 4422 4423 4424 4425 4426
        return NULL;
    }

    return devname;
}

4427 4428
static int qemudDomainChangeEjectableMedia(virConnectPtr conn,
                                           virDomainObjPtr vm,
4429 4430
                                           virDomainDeviceDefPtr dev,
                                           unsigned int qemuCmdFlags)
4431
{
4432
    virDomainDiskDefPtr origdisk = NULL, newdisk;
4433
    char *cmd, *reply, *safe_path;
4434
    char *devname = NULL;
4435
    int i;
4436

4437
    origdisk = NULL;
4438
    newdisk = dev->data.disk;
4439 4440 4441 4442
    for (i = 0 ; i < vm->def->ndisks ; i++) {
        if (vm->def->disks[i]->bus == newdisk->bus &&
            STREQ(vm->def->disks[i]->dst, newdisk->dst)) {
            origdisk = vm->def->disks[i];
4443
            break;
4444
        }
4445 4446 4447
    }

    if (!origdisk) {
4448
        qemudReportError(conn, dom, NULL, VIR_ERR_INTERNAL_ERROR,
4449 4450 4451 4452 4453 4454 4455
                         _("No device with bus '%s' and target '%s'"),
                         virDomainDiskBusTypeToString(newdisk->bus),
                         newdisk->dst);
        return -1;
    }

    if (qemuCmdFlags & QEMUD_CMD_FLAG_DRIVE) {
4456
        if (!(devname = qemudDiskDeviceName(conn, newdisk)))
4457 4458 4459 4460 4461 4462 4463 4464 4465
            return -1;
    } else {
        /* Back compat for no -drive option */
        if (newdisk->device == VIR_DOMAIN_DISK_DEVICE_FLOPPY)
            devname = strdup(newdisk->dst);
        else if (newdisk->device == VIR_DOMAIN_DISK_DEVICE_CDROM &&
                 STREQ(newdisk->dst, "hdc"))
            devname = strdup("cdrom");
        else {
4466
            qemudReportError(conn, dom, NULL, VIR_ERR_INTERNAL_ERROR,
4467 4468 4469 4470 4471 4472 4473 4474
                             _("Emulator version does not support removable "
                               "media for device '%s' and target '%s'"),
                               virDomainDiskDeviceTypeToString(newdisk->device),
                               newdisk->dst);
            return -1;
        }

        if (!devname) {
4475
            virReportOOMError(conn);
4476 4477 4478
            return -1;
        }
    }
4479

4480
    if (newdisk->src) {
4481 4482
        safe_path = qemudEscapeMonitorArg(newdisk->src);
        if (!safe_path) {
4483
            virReportOOMError(conn);
4484
            VIR_FREE(devname);
4485 4486
            return -1;
        }
4487
        if (virAsprintf(&cmd, "change %s \"%s\"", devname, safe_path) == -1) {
4488
            virReportOOMError(conn);
4489
            VIR_FREE(safe_path);
4490
            VIR_FREE(devname);
4491 4492
            return -1;
        }
4493
        VIR_FREE(safe_path);
4494

4495
    } else if (virAsprintf(&cmd, "eject %s", devname) == -1) {
4496
        virReportOOMError(conn);
4497
        VIR_FREE(devname);
4498 4499
        return -1;
    }
4500
    VIR_FREE(devname);
4501

4502
    if (qemudMonitorCommand(vm, cmd, &reply) < 0) {
4503
        qemudReportError(conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
C
Cole Robinson 已提交
4504
                         "%s", _("could not change cdrom media"));
4505
        VIR_FREE(cmd);
4506 4507
        return -1;
    }
4508 4509 4510 4511

    /* If the command failed qemu prints:
     * device not found, device is locked ...
     * No message is printed on success it seems */
4512
    DEBUG ("%s: ejectable media change reply: %s", vm->def->name, reply);
4513
    if (strstr(reply, "\ndevice ")) {
4514
        qemudReportError (conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
C
Cole Robinson 已提交
4515
                          _("changing cdrom media failed: %s"), reply);
4516 4517 4518 4519
        VIR_FREE(reply);
        VIR_FREE(cmd);
        return -1;
    }
4520 4521
    VIR_FREE(reply);
    VIR_FREE(cmd);
4522

4523 4524
    VIR_FREE(origdisk->src);
    origdisk->src = newdisk->src;
4525
    newdisk->src = NULL;
4526
    origdisk->type = newdisk->type;
4527 4528 4529
    return 0;
}

4530 4531 4532 4533 4534 4535 4536 4537 4538 4539 4540 4541 4542 4543 4544 4545 4546 4547 4548 4549 4550 4551 4552 4553 4554 4555 4556 4557 4558 4559 4560 4561 4562 4563 4564 4565 4566 4567 4568 4569 4570 4571 4572 4573 4574 4575 4576 4577 4578 4579 4580 4581 4582 4583 4584 4585 4586 4587 4588 4589 4590 4591 4592 4593 4594 4595 4596
static int
qemudParsePciAddReply(virDomainObjPtr vm,
                      const char *reply,
                      unsigned *domain,
                      unsigned *bus,
                      unsigned *slot)
{
    char *s, *e;

    DEBUG("%s: pci_add reply: %s", vm->def->name, reply);

    /* If the command succeeds qemu prints:
     * OK bus 0, slot XXX...
     * or
     * OK domain 0, bus 0, slot XXX
     */
    if (!(s = strstr(reply, "OK ")))
        return -1;

    s += 3;

    if (STRPREFIX(s, "domain ")) {
        s += strlen("domain ");

        if (virStrToLong_ui(s, &e, 10, domain) == -1) {
            VIR_WARN(_("Unable to parse domain number '%s'\n"), s);
            return -1;
        }

        if (!STRPREFIX(e, ", ")) {
            VIR_WARN(_("Expected ', ' parsing pci_add reply '%s'\n"), s);
            return -1;
        }
        s = e + 2;
    }

    if (!STRPREFIX(s, "bus ")) {
        VIR_WARN(_("Expected 'bus ' parsing pci_add reply '%s'\n"), s);
        return -1;
    }
    s += strlen("bus ");

    if (virStrToLong_ui(s, &e, 10, bus) == -1) {
        VIR_WARN(_("Unable to parse bus number '%s'\n"), s);
        return -1;
    }

    if (!STRPREFIX(e, ", ")) {
        VIR_WARN(_("Expected ', ' parsing pci_add reply '%s'\n"), s);
        return -1;
    }
    s = e + 2;

    if (!STRPREFIX(s, "slot ")) {
        VIR_WARN(_("Expected 'slot ' parsing pci_add reply '%s'\n"), s);
        return -1;
    }
    s += strlen("slot ");

    if (virStrToLong_ui(s, &e, 10, slot) == -1) {
        VIR_WARN(_("Unable to parse slot number '%s'\n"), s);
        return -1;
    }

    return 0;
}

4597 4598 4599
static int qemudDomainAttachPciDiskDevice(virConnectPtr conn,
                                          virDomainObjPtr vm,
                                          virDomainDeviceDefPtr dev)
4600 4601
{
    int ret, i;
4602
    char *cmd, *reply;
4603 4604
    char *safe_path;
    const char* type = virDomainDiskBusTypeToString(dev->data.disk->bus);
4605
    int tryOldSyntax = 0;
4606
    unsigned domain, bus, slot;
4607 4608 4609

    for (i = 0 ; i < vm->def->ndisks ; i++) {
        if (STREQ(vm->def->disks[i]->dst, dev->data.disk->dst)) {
4610
            qemudReportError(conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
4611 4612 4613 4614 4615 4616
                           _("target %s already exists"), dev->data.disk->dst);
            return -1;
        }
    }

    if (VIR_REALLOC_N(vm->def->disks, vm->def->ndisks+1) < 0) {
4617
        virReportOOMError(conn);
4618 4619 4620
        return -1;
    }

4621
try_command:
4622 4623
    safe_path = qemudEscapeMonitorArg(dev->data.disk->src);
    if (!safe_path) {
4624
        virReportOOMError(conn);
4625 4626 4627
        return -1;
    }

4628 4629
    ret = virAsprintf(&cmd, "pci_add %s storage file=%s,if=%s",
                      (tryOldSyntax ? "0": "pci_addr=auto"), safe_path, type);
4630 4631
    VIR_FREE(safe_path);
    if (ret == -1) {
4632
        virReportOOMError(conn);
4633 4634 4635
        return ret;
    }

4636
    if (qemudMonitorCommand(vm, cmd, &reply) < 0) {
4637
        qemudReportError(conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
4638 4639 4640 4641 4642
                         _("cannot attach %s disk"), type);
        VIR_FREE(cmd);
        return -1;
    }

4643 4644
    VIR_FREE(cmd);

4645 4646 4647 4648 4649 4650
    if (qemudParsePciAddReply(vm, reply, &domain, &bus, &slot) < 0) {
        if (!tryOldSyntax && strstr(reply, "invalid char in expression")) {
            VIR_FREE(reply);
            tryOldSyntax = 1;
            goto try_command;
        }
4651

4652
        qemudReportError (conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
4653
                          _("adding %s disk failed: %s"), type, reply);
4654 4655 4656 4657
        VIR_FREE(reply);
        return -1;
    }

4658 4659 4660 4661 4662 4663
    VIR_FREE(reply);

    dev->data.disk->pci_addr.domain = domain;
    dev->data.disk->pci_addr.bus    = bus;
    dev->data.disk->pci_addr.slot   = slot;

4664 4665 4666 4667 4668 4669
    vm->def->disks[vm->def->ndisks++] = dev->data.disk;
    qsort(vm->def->disks, vm->def->ndisks, sizeof(*vm->def->disks),
          virDomainDiskQSort);

    return 0;
}
4670

4671 4672 4673
static int qemudDomainAttachUsbMassstorageDevice(virConnectPtr conn,
                                                 virDomainObjPtr vm,
                                                 virDomainDeviceDefPtr dev)
4674
{
4675 4676
    int ret, i;
    char *safe_path;
4677 4678
    char *cmd, *reply;

4679 4680
    for (i = 0 ; i < vm->def->ndisks ; i++) {
        if (STREQ(vm->def->disks[i]->dst, dev->data.disk->dst)) {
4681
            qemudReportError(conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
4682 4683 4684 4685 4686
                           _("target %s already exists"), dev->data.disk->dst);
            return -1;
        }
    }

4687
    if (VIR_REALLOC_N(vm->def->disks, vm->def->ndisks+1) < 0) {
4688
        virReportOOMError(conn);
4689
        return -1;
4690 4691
    }

4692 4693
    safe_path = qemudEscapeMonitorArg(dev->data.disk->src);
    if (!safe_path) {
4694
        virReportOOMError(conn);
4695 4696 4697
        return -1;
    }

4698
    ret = virAsprintf(&cmd, "usb_add disk:%s", safe_path);
4699
    VIR_FREE(safe_path);
4700
    if (ret == -1) {
4701
        virReportOOMError(conn);
4702 4703 4704
        return ret;
    }

4705
    if (qemudMonitorCommand(vm, cmd, &reply) < 0) {
4706
        qemudReportError(conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
4707
                         "%s", _("cannot attach usb disk"));
4708 4709 4710 4711
        VIR_FREE(cmd);
        return -1;
    }

4712
    DEBUG ("%s: attach_usb reply: %s",vm->def->name,  reply);
4713 4714 4715
    /* If the command failed qemu prints:
     * Could not add ... */
    if (strstr(reply, "Could not add ")) {
4716
        qemudReportError (conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
4717
                          "%s",
4718
                          _("adding usb disk failed"));
4719 4720 4721 4722
        VIR_FREE(reply);
        VIR_FREE(cmd);
        return -1;
    }
4723

4724 4725 4726
    vm->def->disks[vm->def->ndisks++] = dev->data.disk;
    qsort(vm->def->disks, vm->def->ndisks, sizeof(*vm->def->disks),
          virDomainDiskQSort);
4727

4728 4729 4730 4731 4732
    VIR_FREE(reply);
    VIR_FREE(cmd);
    return 0;
}

M
Mark McLoughlin 已提交
4733 4734 4735 4736 4737 4738
static int qemudDomainAttachNetDevice(virConnectPtr conn,
                                      virDomainObjPtr vm,
                                      virDomainDeviceDefPtr dev,
                                      unsigned int qemuCmdFlags)
{
    virDomainNetDefPtr net = dev->data.net;
4739
    char *cmd, *reply, *remove_cmd;
M
Mark McLoughlin 已提交
4740
    int i;
4741
    unsigned domain, bus, slot;
M
Mark McLoughlin 已提交
4742 4743 4744 4745 4746 4747 4748 4749 4750 4751 4752 4753 4754 4755 4756 4757 4758 4759 4760 4761 4762 4763 4764 4765 4766 4767 4768 4769 4770 4771 4772 4773 4774 4775 4776 4777 4778 4779

    if (!(qemuCmdFlags & QEMUD_CMD_FLAG_HOST_NET_ADD)) {
        qemudReportError(conn, dom, NULL, VIR_ERR_NO_SUPPORT, "%s",
                         _("installed qemu version does not support host_net_add"));
        return -1;
    }

    if (net->type == VIR_DOMAIN_NET_TYPE_BRIDGE ||
        net->type == VIR_DOMAIN_NET_TYPE_NETWORK) {
        qemudReportError(conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                         _("network device type '%s' cannot be attached"),
                         virDomainNetTypeToString(net->type));
        return -1;
    }

    if (VIR_REALLOC_N(vm->def->nets, vm->def->nnets+1) < 0) {
        virReportOOMError(conn);
        return -1;
    }

    if ((qemuCmdFlags & QEMUD_CMD_FLAG_NET_NAME) &&
        qemuAssignNetNames(vm->def, net) < 0) {
        virReportOOMError(conn);
        return -1;
    }

    /* Choose a vlan value greater than all other values since
     * older versions did not store the value in the state file.
     */
    net->vlan = vm->def->nnets;
    for (i = 0; i < vm->def->nnets; i++)
        if (vm->def->nets[i]->vlan >= net->vlan)
            net->vlan = vm->def->nets[i]->vlan;

    if (qemuBuildHostNetStr(conn, net,
                            "host_net_add ", ' ', net->vlan, -1, &cmd) < 0)
        return -1;

4780 4781 4782 4783 4784 4785 4786 4787 4788
    remove_cmd = NULL;
    if (net->vlan >= 0 && net->hostnet_name &&
        virAsprintf(&remove_cmd, "host_net_remove %d %s",
                    net->vlan, net->hostnet_name) < 0) {
        VIR_FREE(cmd);
        virReportOOMError(conn);
        return -1;
    }

M
Mark McLoughlin 已提交
4789 4790 4791
    if (qemudMonitorCommand(vm, cmd, &reply) < 0) {
        qemudReportError(conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                         _("failed to add network backend with '%s'"), cmd);
4792
        VIR_FREE(remove_cmd);
M
Mark McLoughlin 已提交
4793 4794 4795 4796 4797 4798 4799 4800 4801 4802
        VIR_FREE(cmd);
        return -1;
    }

    DEBUG("%s: host_net_add reply: %s", vm->def->name, reply);

    VIR_FREE(reply);
    VIR_FREE(cmd);

    if (qemuBuildNicStr(conn, net,
4803 4804
                        "pci_add pci_addr=auto ", ' ', net->vlan, &cmd) < 0)
        goto try_remove;
M
Mark McLoughlin 已提交
4805 4806 4807 4808 4809

    if (qemudMonitorCommand(vm, cmd, &reply) < 0) {
        qemudReportError(conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                         _("failed to add NIC with '%s'"), cmd);
        VIR_FREE(cmd);
4810
        goto try_remove;
M
Mark McLoughlin 已提交
4811 4812 4813
    }

    VIR_FREE(cmd);
4814 4815 4816 4817 4818 4819 4820 4821 4822

    if (qemudParsePciAddReply(vm, reply, &domain, &bus, &slot) < 0) {
        qemudReportError(conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                         _("parsing pci_add reply failed: %s"), reply);
        VIR_FREE(reply);
        goto try_remove;
    }

    VIR_FREE(reply);
4823
    VIR_FREE(remove_cmd);
M
Mark McLoughlin 已提交
4824

4825 4826 4827 4828
    net->pci_addr.domain = domain;
    net->pci_addr.bus    = bus;
    net->pci_addr.slot   = slot;

M
Mark McLoughlin 已提交
4829 4830 4831
    vm->def->nets[vm->def->nnets++] = net;

    return 0;
4832 4833 4834 4835 4836 4837 4838 4839 4840 4841 4842 4843 4844 4845 4846

try_remove:
    reply = NULL;

    if (!remove_cmd)
        VIR_WARN0(_("Unable to remove network backend\n"));
    else if (qemudMonitorCommand(vm, remove_cmd, &reply) < 0)
        VIR_WARN(_("Failed to remove network backend with '%s'\n"), remove_cmd);
    else
        VIR_DEBUG("%s: host_net_remove reply: %s\n", vm->def->name, reply);

    VIR_FREE(reply);
    VIR_FREE(remove_cmd);

    return -1;
M
Mark McLoughlin 已提交
4847 4848
}

4849 4850 4851
static int qemudDomainAttachHostDevice(virConnectPtr conn,
                                       virDomainObjPtr vm,
                                       virDomainDeviceDefPtr dev)
4852 4853 4854 4855
{
    int ret;
    char *cmd, *reply;

4856
    if (VIR_REALLOC_N(vm->def->hostdevs, vm->def->nhostdevs+1) < 0) {
4857
        virReportOOMError(conn);
4858 4859
        return -1;
    }
4860

4861
    if (dev->data.hostdev->source.subsys.u.usb.vendor) {
4862 4863 4864
        ret = virAsprintf(&cmd, "usb_add host:%.4x:%.4x",
                          dev->data.hostdev->source.subsys.u.usb.vendor,
                          dev->data.hostdev->source.subsys.u.usb.product);
4865
    } else {
4866 4867 4868
        ret = virAsprintf(&cmd, "usb_add host:%.3d.%.3d",
                          dev->data.hostdev->source.subsys.u.usb.bus,
                          dev->data.hostdev->source.subsys.u.usb.device);
4869 4870
    }
    if (ret == -1) {
4871
        virReportOOMError(conn);
4872 4873 4874
        return -1;
    }

4875
    if (qemudMonitorCommand(vm, cmd, &reply) < 0) {
4876
        qemudReportError(conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
4877 4878 4879 4880 4881
                         "%s", _("cannot attach usb device"));
        VIR_FREE(cmd);
        return -1;
    }

4882
    DEBUG ("%s: attach_usb reply: %s", vm->def->name, reply);
4883 4884 4885
    /* If the command failed qemu prints:
     * Could not add ... */
    if (strstr(reply, "Could not add ")) {
4886
        qemudReportError (conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
4887 4888 4889 4890 4891 4892
                          "%s",
                          _("adding usb device failed"));
        VIR_FREE(reply);
        VIR_FREE(cmd);
        return -1;
    }
4893

4894
    vm->def->hostdevs[vm->def->nhostdevs++] = dev->data.hostdev;
4895

4896 4897
    VIR_FREE(reply);
    VIR_FREE(cmd);
4898 4899 4900
    return 0;
}

4901 4902
static int qemudDomainAttachDevice(virDomainPtr dom,
                                   const char *xml) {
4903 4904 4905
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    virDomainDeviceDefPtr dev = NULL;
4906
    unsigned int qemuCmdFlags;
4907
    virCgroupPtr cgroup = NULL;
4908
    int ret = -1;
4909

4910
    qemuDriverLock(driver);
4911
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
4912
    if (!vm) {
4913 4914 4915 4916
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
4917
        goto cleanup;
4918 4919 4920
    }

    if (!virDomainIsActive(vm)) {
4921
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
4922
                         "%s", _("cannot attach device on inactive domain"));
4923
        goto cleanup;
4924 4925
    }

4926 4927
    dev = virDomainDeviceDefParse(dom->conn, driver->caps, vm->def, xml,
                                  VIR_DOMAIN_XML_INACTIVE);
4928 4929 4930
    if (dev == NULL)
        goto cleanup;

4931 4932 4933 4934
    if (qemudExtractVersionInfo(vm->def->emulator,
                                NULL,
                                &qemuCmdFlags) < 0)
        goto cleanup;
4935

4936
    if (dev->type == VIR_DOMAIN_DEVICE_DISK) {
4937 4938 4939 4940 4941 4942 4943 4944 4945 4946 4947 4948 4949 4950 4951 4952 4953 4954 4955 4956 4957
        if (driver->cgroup != NULL) {
            if (virCgroupForDomain(driver->cgroup, vm->def->name, &cgroup, 0) !=0 ) {
                qemudReportError(dom->conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                                 _("Unable to find cgroup for %s\n"),
                                 vm->def->name);
                goto cleanup;
            }
            if (dev->data.disk->src != NULL &&
                dev->data.disk->type == VIR_DOMAIN_DISK_TYPE_BLOCK &&
                virCgroupAllowDevicePath(cgroup,
                                         dev->data.disk->src) < 0) {
                qemudReportError(dom->conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                                 _("unable to allow device %s"),
                                 dev->data.disk->src);
                goto cleanup;
            }
        }

        if (driver->securityDriver)
            driver->securityDriver->domainSetSecurityImageLabel(dom->conn, vm, dev->data.disk);

4958
        switch (dev->data.disk->device) {
4959 4960
        case VIR_DOMAIN_DISK_DEVICE_CDROM:
        case VIR_DOMAIN_DISK_DEVICE_FLOPPY:
4961 4962
            if (driver->securityDriver)
                driver->securityDriver->domainSetSecurityImageLabel(dom->conn, vm, dev->data.disk);
4963 4964 4965 4966

            if (qemuDomainSetDeviceOwnership(dom->conn, driver, dev, 0) < 0)
                goto cleanup;

4967
            ret = qemudDomainChangeEjectableMedia(dom->conn, vm, dev, qemuCmdFlags);
4968
            break;
4969

4970
        case VIR_DOMAIN_DISK_DEVICE_DISK:
4971 4972
            if (driver->securityDriver)
                driver->securityDriver->domainSetSecurityImageLabel(dom->conn, vm, dev->data.disk);
4973 4974 4975 4976

            if (qemuDomainSetDeviceOwnership(dom->conn, driver, dev, 0) < 0)
                goto cleanup;

4977
            if (dev->data.disk->bus == VIR_DOMAIN_DISK_BUS_USB) {
4978
                ret = qemudDomainAttachUsbMassstorageDevice(dom->conn, vm, dev);
4979 4980
            } else if (dev->data.disk->bus == VIR_DOMAIN_DISK_BUS_SCSI ||
                       dev->data.disk->bus == VIR_DOMAIN_DISK_BUS_VIRTIO) {
4981
                ret = qemudDomainAttachPciDiskDevice(dom->conn, vm, dev);
4982 4983 4984 4985
            } else {
                qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                                 _("disk bus '%s' cannot be hotplugged."),
                                 virDomainDiskBusTypeToString(dev->data.disk->bus));
4986
                /* fallthrough */
4987 4988
            }
            break;
4989

4990 4991
        default:
            qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
4992 4993
                             _("disk device type '%s' cannot be hotplugged"),
                             virDomainDiskDeviceTypeToString(dev->data.disk->device));
4994 4995 4996 4997 4998
            /* Fallthrough */
        }
        if (ret != 0) {
            virCgroupDenyDevicePath(cgroup,
                                    dev->data.disk->src);
4999
        }
M
Mark McLoughlin 已提交
5000 5001
    } else if (dev->type == VIR_DOMAIN_DEVICE_NET) {
        ret = qemudDomainAttachNetDevice(dom->conn, vm, dev, qemuCmdFlags);
5002
    } else if (dev->type == VIR_DOMAIN_DEVICE_HOSTDEV &&
5003 5004
               dev->data.hostdev->mode == VIR_DOMAIN_HOSTDEV_MODE_SUBSYS &&
               dev->data.hostdev->source.subsys.type == VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_USB) {
5005 5006 5007
        if (qemuDomainSetDeviceOwnership(dom->conn, driver, dev, 0) < 0)
            goto cleanup;

5008
        ret = qemudDomainAttachHostDevice(dom->conn, vm, dev);
5009
    } else {
5010
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
5011 5012
                         _("device type '%s' cannot be attached"),
                         virDomainDeviceTypeToString(dev->type));
5013
        goto cleanup;
5014 5015
    }

5016
    if (!ret && virDomainSaveStatus(dom->conn, driver->stateDir, vm) < 0)
5017 5018
        ret = -1;

5019
cleanup:
5020 5021 5022
    if (cgroup)
        virCgroupFree(&cgroup);

5023 5024 5025
    if (ret < 0) {
        if (qemuDomainSetDeviceOwnership(dom->conn, driver, dev, 1) < 0)
            VIR_WARN0("Fail to restore disk device ownership");
G
Guido Günther 已提交
5026
        virDomainDeviceDefFree(dev);
5027
    }
5028 5029
    if (vm)
        virDomainObjUnlock(vm);
5030
    qemuDriverUnlock(driver);
5031 5032 5033
    return ret;
}

5034 5035
static int qemudDomainDetachPciDiskDevice(virConnectPtr conn,
                                          virDomainObjPtr vm, virDomainDeviceDefPtr dev)
5036 5037
{
    int i, ret = -1;
5038 5039
    char *cmd = NULL;
    char *reply = NULL;
5040
    virDomainDiskDefPtr detach = NULL;
5041
    int tryOldSyntax = 0;
5042 5043 5044 5045 5046 5047 5048 5049 5050

    for (i = 0 ; i < vm->def->ndisks ; i++) {
        if (STREQ(vm->def->disks[i]->dst, dev->data.disk->dst)) {
            detach = vm->def->disks[i];
            break;
        }
    }

    if (!detach) {
5051
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
5052
                         _("disk %s not found"), dev->data.disk->dst);
5053
        goto cleanup;
5054 5055
    }

5056
    if (!virDiskHasValidPciAddr(detach)) {
5057
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
5058 5059
                         _("disk %s cannot be detached - no PCI address for device"),
                           detach->dst);
5060
        goto cleanup;
5061 5062
    }

5063 5064
try_command:
    if (tryOldSyntax) {
5065
        if (virAsprintf(&cmd, "pci_del 0 %.2x", detach->pci_addr.slot) < 0) {
5066 5067 5068 5069
            virReportOOMError(conn);
            goto cleanup;
        }
    } else {
5070 5071 5072 5073
        if (virAsprintf(&cmd, "pci_del pci_addr=%.4x:%.2x:%.2x",
                        detach->pci_addr.domain,
                        detach->pci_addr.bus,
                        detach->pci_addr.slot) < 0) {
5074 5075 5076
            virReportOOMError(conn);
            goto cleanup;
        }
5077 5078
    }

5079
    if (qemudMonitorCommand(vm, cmd, &reply) < 0) {
5080
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
5081
                          _("failed to execute detach disk %s command"), detach->dst);
5082
        goto cleanup;
5083 5084
    }

5085
    DEBUG ("%s: pci_del reply: %s",vm->def->name,  reply);
5086 5087 5088 5089 5090 5091

    if (!tryOldSyntax &&
        strstr(reply, "extraneous characters")) {
        tryOldSyntax = 1;
        goto try_command;
    }
5092 5093
    /* If the command fails due to a wrong slot qemu prints: invalid slot,
     * nothing is printed on success */
5094 5095
    if (strstr(reply, "invalid slot") ||
        strstr(reply, "Invalid pci address")) {
5096
        qemudReportError (conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
5097 5098 5099 5100 5101 5102
                          _("failed to detach disk %s: invalid PCI address %.4x:%.2x:%.2x: %s"),
                          detach->dst,
                          detach->pci_addr.domain,
                          detach->pci_addr.bus,
                          detach->pci_addr.slot,
                          reply);
5103
        goto cleanup;
5104 5105 5106 5107 5108
    }

    if (vm->def->ndisks > 1) {
        vm->def->disks[i] = vm->def->disks[--vm->def->ndisks];
        if (VIR_REALLOC_N(vm->def->disks, vm->def->ndisks) < 0) {
5109
            virReportOOMError(conn);
5110
            goto cleanup;
5111 5112 5113 5114 5115 5116 5117 5118
        }
        qsort(vm->def->disks, vm->def->ndisks, sizeof(*vm->def->disks),
              virDomainDiskQSort);
    } else {
        VIR_FREE(vm->def->disks[0]);
        vm->def->ndisks = 0;
    }
    ret = 0;
5119 5120

cleanup:
5121 5122 5123 5124 5125
    VIR_FREE(reply);
    VIR_FREE(cmd);
    return ret;
}

5126 5127 5128 5129 5130 5131 5132 5133 5134 5135 5136 5137 5138 5139 5140 5141 5142 5143 5144 5145 5146 5147 5148 5149 5150 5151 5152 5153 5154 5155 5156 5157 5158 5159 5160 5161 5162 5163 5164 5165 5166 5167 5168 5169 5170 5171 5172 5173 5174 5175 5176 5177 5178 5179 5180 5181 5182 5183 5184 5185 5186 5187 5188 5189 5190 5191 5192 5193 5194 5195 5196 5197 5198 5199 5200 5201 5202 5203 5204 5205 5206 5207 5208 5209 5210 5211 5212 5213 5214 5215 5216 5217 5218 5219 5220 5221 5222
static int
qemudDomainDetachNetDevice(virConnectPtr conn,
                           virDomainObjPtr vm,
                           virDomainDeviceDefPtr dev)
{
    int i, ret = -1;
    char *cmd = NULL;
    char *reply = NULL;
    virDomainNetDefPtr detach = NULL;

    for (i = 0 ; i < vm->def->nnets ; i++) {
        virDomainNetDefPtr net = vm->def->nets[i];

        if (!memcmp(net->mac, dev->data.net->mac,  sizeof(net->mac))) {
            detach = net;
            break;
        }
    }

    if (!detach) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                         _("network device %02x:%02x:%02x:%02x:%02x:%02x not found"),
                         dev->data.net->mac[0], dev->data.net->mac[1],
                         dev->data.net->mac[2], dev->data.net->mac[3],
                         dev->data.net->mac[4], dev->data.net->mac[5]);
        goto cleanup;
    }

    if (!virNetHasValidPciAddr(detach) || detach->vlan < 0 || !detach->hostnet_name) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                         "%s", _("network device cannot be detached - device state missing"));
        goto cleanup;
    }

    if (virAsprintf(&cmd, "pci_del pci_addr=%.4x:%.2x:%.2x",
                    detach->pci_addr.domain,
                    detach->pci_addr.bus,
                    detach->pci_addr.slot) < 0) {
        virReportOOMError(conn);
        goto cleanup;
    }

    if (qemudMonitorCommand(vm, cmd, &reply) < 0) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                          _("network device dettach command '%s' failed"), cmd);
        goto cleanup;
    }

    DEBUG("%s: pci_del reply: %s", vm->def->name,  reply);

    /* If the command fails due to a wrong PCI address qemu prints
     * 'invalid pci address'; nothing is printed on success */
    if (strstr(reply, "Invalid pci address")) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                         _("failed to detach network device: invalid PCI address %.4x:%.2x:%.2x: %s"),
                         detach->pci_addr.domain,
                         detach->pci_addr.bus,
                         detach->pci_addr.slot,
                         reply);
        goto cleanup;
    }

    VIR_FREE(reply);
    VIR_FREE(cmd);

    if (virAsprintf(&cmd, "host_net_remove %d %s",
                    detach->vlan, detach->hostnet_name) < 0) {
        virReportOOMError(conn);
        goto cleanup;
    }

    if (qemudMonitorCommand(vm, cmd, &reply) < 0) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                          _("network device dettach command '%s' failed"), cmd);
        goto cleanup;
    }

    DEBUG("%s: host_net_remove reply: %s", vm->def->name,  reply);

    if (vm->def->nnets > 1) {
        vm->def->nets[i] = vm->def->nets[--vm->def->nnets];
        if (VIR_REALLOC_N(vm->def->nets, vm->def->nnets) < 0) {
            virReportOOMError(conn);
            goto cleanup;
        }
    } else {
        VIR_FREE(vm->def->nets[0]);
        vm->def->nnets = 0;
    }
    ret = 0;

cleanup:
    VIR_FREE(reply);
    VIR_FREE(cmd);
    return ret;
}

5223 5224
static int qemudDomainDetachDevice(virDomainPtr dom,
                                   const char *xml) {
5225 5226 5227 5228
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    virDomainDeviceDefPtr dev = NULL;
    int ret = -1;
5229

5230
    qemuDriverLock(driver);
5231
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
5232
    if (!vm) {
5233 5234 5235 5236
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
5237
        goto cleanup;
5238 5239 5240
    }

    if (!virDomainIsActive(vm)) {
5241
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
G
Guido Günther 已提交
5242
                         "%s", _("cannot detach device on inactive domain"));
5243
        goto cleanup;
5244 5245
    }

5246 5247
    dev = virDomainDeviceDefParse(dom->conn, driver->caps, vm->def, xml,
                                  VIR_DOMAIN_XML_INACTIVE);
5248 5249 5250
    if (dev == NULL)
        goto cleanup;

5251 5252 5253 5254

    if (dev->type == VIR_DOMAIN_DEVICE_DISK &&
        dev->data.disk->device == VIR_DOMAIN_DISK_DEVICE_DISK &&
        (dev->data.disk->bus == VIR_DOMAIN_DISK_BUS_SCSI ||
5255
         dev->data.disk->bus == VIR_DOMAIN_DISK_BUS_VIRTIO)) {
5256
        ret = qemudDomainDetachPciDiskDevice(dom->conn, vm, dev);
5257
        if (driver->securityDriver)
5258
            driver->securityDriver->domainRestoreSecurityImageLabel(dom->conn, dev->data.disk);
5259 5260
        if (qemuDomainSetDeviceOwnership(dom->conn, driver, dev, 1) < 0)
            VIR_WARN0("Fail to restore disk device ownership");
5261 5262 5263
    } else if (dev->type == VIR_DOMAIN_DEVICE_NET) {
        ret = qemudDomainDetachNetDevice(dom->conn, vm, dev);
    } else
5264 5265 5266
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                         "%s", _("only SCSI or virtio disk device can be detached dynamically"));

5267
    if (!ret && virDomainSaveStatus(dom->conn, driver->stateDir, vm) < 0)
5268 5269
        ret = -1;

5270 5271
cleanup:
    virDomainDeviceDefFree(dev);
5272 5273
    if (vm)
        virDomainObjUnlock(vm);
5274
    qemuDriverUnlock(driver);
5275 5276 5277
    return ret;
}

5278
static int qemudDomainGetAutostart(virDomainPtr dom,
5279
                                   int *autostart) {
5280 5281 5282
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    int ret = -1;
5283

5284
    qemuDriverLock(driver);
5285
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
5286 5287
    qemuDriverUnlock(driver);

5288
    if (!vm) {
5289 5290 5291 5292
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
5293
        goto cleanup;
5294 5295 5296
    }

    *autostart = vm->autostart;
5297
    ret = 0;
5298

5299
cleanup:
5300 5301
    if (vm)
        virDomainObjUnlock(vm);
5302
    return ret;
5303 5304
}

5305
static int qemudDomainSetAutostart(virDomainPtr dom,
5306
                                   int autostart) {
5307 5308
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
5309 5310
    char *configFile = NULL, *autostartLink = NULL;
    int ret = -1;
5311

5312
    qemuDriverLock(driver);
5313
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
5314

5315
    if (!vm) {
5316 5317 5318 5319
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
5320
        goto cleanup;
5321 5322
    }

5323 5324 5325
    if (!vm->persistent) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INTERNAL_ERROR,
                         "%s", _("cannot set autostart for transient domain"));
5326
        goto cleanup;
5327 5328
    }

5329 5330
    autostart = (autostart != 0);

5331 5332 5333 5334 5335
    if (vm->autostart != autostart) {
        if ((configFile = virDomainConfigFile(dom->conn, driver->configDir, vm->def->name)) == NULL)
            goto cleanup;
        if ((autostartLink = virDomainConfigFile(dom->conn, driver->autostartDir, vm->def->name)) == NULL)
            goto cleanup;
5336

5337 5338
        if (autostart) {
            int err;
5339

5340
            if ((err = virFileMakePath(driver->autostartDir))) {
5341 5342 5343
                virReportSystemError(dom->conn, err,
                                     _("cannot create autostart directory %s"),
                                     driver->autostartDir);
5344 5345
                goto cleanup;
            }
5346

5347
            if (symlink(configFile, autostartLink) < 0) {
5348 5349 5350
                virReportSystemError(dom->conn, errno,
                                     _("Failed to create symlink '%s to '%s'"),
                                     autostartLink, configFile);
5351 5352 5353 5354
                goto cleanup;
            }
        } else {
            if (unlink(autostartLink) < 0 && errno != ENOENT && errno != ENOTDIR) {
5355 5356 5357
                virReportSystemError(dom->conn, errno,
                                     _("Failed to delete symlink '%s'"),
                                     autostartLink);
5358 5359
                goto cleanup;
            }
5360 5361
        }

5362
        vm->autostart = autostart;
5363
    }
5364
    ret = 0;
5365

5366 5367 5368
cleanup:
    VIR_FREE(configFile);
    VIR_FREE(autostartLink);
5369 5370
    if (vm)
        virDomainObjUnlock(vm);
5371
    qemuDriverUnlock(driver);
5372
    return ret;
5373 5374
}

5375 5376 5377 5378 5379 5380 5381 5382 5383 5384 5385 5386 5387 5388 5389 5390 5391 5392 5393 5394 5395 5396 5397 5398 5399 5400 5401 5402 5403 5404 5405 5406 5407 5408 5409 5410 5411 5412 5413 5414 5415 5416 5417 5418 5419 5420 5421 5422 5423 5424 5425 5426 5427 5428 5429 5430 5431 5432 5433 5434 5435 5436 5437 5438 5439 5440 5441 5442 5443 5444 5445 5446 5447 5448 5449 5450 5451 5452 5453 5454 5455 5456 5457 5458 5459 5460 5461 5462 5463 5464 5465 5466 5467 5468 5469 5470 5471 5472 5473 5474 5475 5476 5477 5478 5479 5480 5481 5482 5483 5484 5485 5486 5487 5488 5489 5490 5491 5492 5493 5494 5495 5496 5497 5498 5499 5500 5501 5502 5503 5504 5505 5506 5507 5508 5509 5510 5511 5512 5513 5514 5515 5516 5517 5518 5519

static char *qemuGetSchedulerType(virDomainPtr dom,
                                  int *nparams)
{
    struct qemud_driver *driver = dom->conn->privateData;
    char *ret;

    if (driver->cgroup == NULL) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                         __FUNCTION__);
        return NULL;
    }

    if (nparams)
        *nparams = 1;

    ret = strdup("posix");
    if (!ret)
        virReportOOMError(dom->conn);
    return ret;
}

static int qemuSetSchedulerParameters(virDomainPtr dom,
                                      virSchedParameterPtr params,
                                      int nparams)
{
    struct qemud_driver *driver = dom->conn->privateData;
    int i;
    virCgroupPtr group = NULL;
    virDomainObjPtr vm = NULL;
    int ret = -1;

    if (driver->cgroup == NULL) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                         __FUNCTION__);
        return -1;
    }

    qemuDriverLock(driver);
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
    qemuDriverUnlock(driver);

    if (vm == NULL) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("No such domain %s"), dom->uuid);
        goto cleanup;
    }

    if (virCgroupForDomain(driver->cgroup, vm->def->name, &group, 0) != 0) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("cannot find cgroup for domain %s"), vm->def->name);
        goto cleanup;
    }

    for (i = 0; i < nparams; i++) {
        virSchedParameterPtr param = &params[i];

        if (STREQ(param->field, "cpu_shares")) {
            int rc;
            if (param->type != VIR_DOMAIN_SCHED_FIELD_ULLONG) {
                qemudReportError(dom->conn, dom, NULL, VIR_ERR_INVALID_ARG,
                                 _("invalid type for cpu_shares tunable, expected a 'ullong'"));
                goto cleanup;
            }

            rc = virCgroupSetCpuShares(group, params[i].value.ul);
            if (rc != 0) {
                virReportSystemError(dom->conn, -rc, "%s",
                                     _("unable to set cpu shares tunable"));
                goto cleanup;
            }
        } else {
            qemudReportError(dom->conn, domain, NULL, VIR_ERR_INVALID_ARG,
                             _("Invalid parameter `%s'"), param->field);
            goto cleanup;
        }
    }
    ret = 0;

cleanup:
    virCgroupFree(&group);
    if (vm)
        virDomainObjUnlock(vm);
    return ret;
}

static int qemuGetSchedulerParameters(virDomainPtr dom,
                                      virSchedParameterPtr params,
                                      int *nparams)
{
    struct qemud_driver *driver = dom->conn->privateData;
    virCgroupPtr group = NULL;
    virDomainObjPtr vm = NULL;
    unsigned long long val;
    int ret = -1;
    int rc;

    if (driver->cgroup == NULL) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                         __FUNCTION__);
        return -1;
    }

    if ((*nparams) != 1) {
        qemudReportError(dom->conn, domain, NULL, VIR_ERR_INVALID_ARG,
                         "%s", _("Invalid parameter count"));
        return -1;
    }

    qemuDriverLock(driver);
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
    qemuDriverUnlock(driver);

    if (vm == NULL) {
        qemudReportError(dom->conn, domain, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("No such domain %s"), dom->uuid);
        goto cleanup;
    }

    if (virCgroupForDomain(driver->cgroup, vm->def->name, &group, 0) != 0) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("cannot find cgroup for domain %s"), vm->def->name);
        goto cleanup;
    }

    rc = virCgroupGetCpuShares(group, &val);
    if (rc != 0) {
        virReportSystemError(dom->conn, -rc, "%s",
                             _("unable to get cpu shares tunable"));
        goto cleanup;
    }
    params[0].value.ul = val;
    strncpy(params[0].field, "cpu_shares", sizeof(params[0].field));
    params[0].type = VIR_DOMAIN_SCHED_FIELD_ULLONG;

    ret = 0;

cleanup:
    virCgroupFree(&group);
    if (vm)
        virDomainObjUnlock(vm);
    return ret;
}


5520 5521 5522 5523 5524 5525 5526 5527 5528
/* This uses the 'info blockstats' monitor command which was
 * integrated into both qemu & kvm in late 2007.  If the command is
 * not supported we detect this and return the appropriate error.
 */
static int
qemudDomainBlockStats (virDomainPtr dom,
                       const char *path,
                       struct _virDomainBlockStats *stats)
{
5529
    struct qemud_driver *driver = dom->conn->privateData;
5530
    char *dummy, *info = NULL;
5531
    const char *p, *eol;
5532
    const char *qemu_dev_name = NULL;
5533
    size_t len;
5534
    int i, ret = -1;
5535
    virDomainObjPtr vm;
5536
    virDomainDiskDefPtr disk = NULL;
5537

5538
    qemuDriverLock(driver);
5539
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
5540
    qemuDriverUnlock(driver);
5541
    if (!vm) {
5542 5543 5544 5545
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
5546
        goto cleanup;
5547
    }
5548
    if (!virDomainIsActive (vm)) {
5549
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
5550
                          "%s", _("domain is not running"));
5551
        goto cleanup;
5552 5553
    }

5554 5555 5556 5557 5558 5559 5560 5561
    for (i = 0 ; i < vm->def->ndisks ; i++) {
        if (STREQ(path, vm->def->disks[i]->dst)) {
            disk = vm->def->disks[i];
            break;
        }
    }

    if (!disk) {
5562 5563
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_INVALID_ARG,
                          _("invalid path: %s"), path);
5564
        goto cleanup;
5565 5566
    }

5567
    qemu_dev_name = qemudDiskDeviceName(dom->conn, disk);
5568
    if (!qemu_dev_name)
5569
        goto cleanup;
5570 5571
    len = strlen (qemu_dev_name);

5572
    if (qemudMonitorCommand (vm, "info blockstats", &info) < 0) {
5573 5574
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                          "%s", _("'info blockstats' command failed"));
5575
        goto cleanup;
5576
    }
5577
    DEBUG ("%s: info blockstats reply: %s", vm->def->name, info);
5578 5579 5580 5581 5582 5583

    /* If the command isn't supported then qemu prints the supported
     * info commands, so the output starts "info ".  Since this is
     * unlikely to be the name of a block device, we can use this
     * to detect if qemu supports the command.
     */
C
Cole Robinson 已提交
5584
    if (strstr(info, "\ninfo ")) {
5585 5586 5587
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                          "%s",
                          _("'info blockstats' not supported by this qemu"));
5588
        goto cleanup;
5589 5590 5591 5592 5593 5594 5595 5596 5597 5598 5599 5600 5601 5602 5603 5604 5605 5606 5607 5608 5609 5610 5611 5612 5613 5614
    }

    stats->rd_req = -1;
    stats->rd_bytes = -1;
    stats->wr_req = -1;
    stats->wr_bytes = -1;
    stats->errs = -1;

    /* The output format for both qemu & KVM is:
     *   blockdevice: rd_bytes=% wr_bytes=% rd_operations=% wr_operations=%
     *   (repeated for each block device)
     * where '%' is a 64 bit number.
     */
    p = info;

    while (*p) {
        if (STREQLEN (p, qemu_dev_name, len)
            && p[len] == ':' && p[len+1] == ' ') {

            eol = strchr (p, '\n');
            if (!eol)
                eol = p + strlen (p);

            p += len+2;         /* Skip to first label. */

            while (*p) {
5615
                if (STRPREFIX (p, "rd_bytes=")) {
5616 5617
                    p += 9;
                    if (virStrToLong_ll (p, &dummy, 10, &stats->rd_bytes) == -1)
5618 5619
                        DEBUG ("%s: error reading rd_bytes: %s",
                               vm->def->name, p);
5620
                } else if (STRPREFIX (p, "wr_bytes=")) {
5621 5622
                    p += 9;
                    if (virStrToLong_ll (p, &dummy, 10, &stats->wr_bytes) == -1)
5623 5624
                        DEBUG ("%s: error reading wr_bytes: %s",
                               vm->def->name, p);
5625
                } else if (STRPREFIX (p, "rd_operations=")) {
5626 5627
                    p += 14;
                    if (virStrToLong_ll (p, &dummy, 10, &stats->rd_req) == -1)
5628 5629
                        DEBUG ("%s: error reading rd_req: %s",
                               vm->def->name, p);
5630
                } else if (STRPREFIX (p, "wr_operations=")) {
5631 5632
                    p += 14;
                    if (virStrToLong_ll (p, &dummy, 10, &stats->wr_req) == -1)
5633 5634
                        DEBUG ("%s: error reading wr_req: %s",
                               vm->def->name, p);
5635
                } else
5636
                    DEBUG ("%s: unknown block stat near %s", vm->def->name, p);
5637 5638 5639 5640 5641 5642

                /* Skip to next label. */
                p = strchr (p, ' ');
                if (!p || p >= eol) break;
                p++;
            }
5643
            ret = 0;
5644
            goto cleanup;
5645 5646 5647 5648 5649 5650 5651 5652 5653 5654 5655
        }

        /* Skip to next line. */
        p = strchr (p, '\n');
        if (!p) break;
        p++;
    }

    /* If we reach here then the device was not found. */
    qemudReportError (dom->conn, dom, NULL, VIR_ERR_INVALID_ARG,
                      _("device not found: %s (%s)"), path, qemu_dev_name);
5656
 cleanup:
5657 5658
    VIR_FREE(qemu_dev_name);
    VIR_FREE(info);
5659 5660
    if (vm)
        virDomainObjUnlock(vm);
5661
    return ret;
5662 5663
}

5664
#ifdef __linux__
5665 5666 5667 5668 5669
static int
qemudDomainInterfaceStats (virDomainPtr dom,
                           const char *path,
                           struct _virDomainInterfaceStats *stats)
{
5670 5671
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
5672
    int i;
5673
    int ret = -1;
5674

5675
    qemuDriverLock(driver);
5676
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
5677 5678
    qemuDriverUnlock(driver);

5679
    if (!vm) {
5680 5681 5682 5683
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
5684
        goto cleanup;
5685 5686
    }

5687
    if (!virDomainIsActive(vm)) {
5688
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
5689
                         "%s", _("domain is not running"));
5690
        goto cleanup;
5691 5692 5693 5694
    }

    if (!path || path[0] == '\0') {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INVALID_ARG,
5695
                         "%s", _("NULL or empty path"));
5696
        goto cleanup;
5697 5698 5699
    }

    /* Check the path is one of the domain's network interfaces. */
5700 5701
    for (i = 0 ; i < vm->def->nnets ; i++) {
        if (vm->def->nets[i]->ifname &&
5702 5703 5704 5705
            STREQ (vm->def->nets[i]->ifname, path)) {
            ret = 0;
            break;
        }
5706 5707
    }

5708 5709 5710 5711 5712
    if (ret == 0)
        ret = linuxDomainInterfaceStats (dom->conn, path, stats);
    else
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_INVALID_ARG,
                          _("invalid path, '%s' is not a known interface"), path);
5713

5714
cleanup:
5715 5716
    if (vm)
        virDomainObjUnlock(vm);
5717 5718
    return ret;
}
5719
#else
5720 5721 5722 5723
static int
qemudDomainInterfaceStats (virDomainPtr dom,
                           const char *path ATTRIBUTE_UNUSED,
                           struct _virDomainInterfaceStats *stats ATTRIBUTE_UNUSED)
5724 5725 5726 5727
    qemudReportError (dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                      "%s", __FUNCTION__);
    return -1;
}
5728
#endif
5729

5730 5731 5732 5733 5734 5735 5736
static int
qemudDomainBlockPeek (virDomainPtr dom,
                      const char *path,
                      unsigned long long offset, size_t size,
                      void *buffer,
                      unsigned int flags ATTRIBUTE_UNUSED)
{
5737 5738 5739
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    int fd = -1, ret = -1, i;
5740

5741
    qemuDriverLock(driver);
5742
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
5743 5744
    qemuDriverUnlock(driver);

5745
    if (!vm) {
5746 5747 5748 5749
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
5750
        goto cleanup;
5751 5752 5753 5754
    }

    if (!path || path[0] == '\0') {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INVALID_ARG,
J
Jim Meyering 已提交
5755
                         "%s", _("NULL or empty path"));
5756
        goto cleanup;
5757 5758 5759
    }

    /* Check the path belongs to this domain. */
5760 5761
    for (i = 0 ; i < vm->def->ndisks ; i++) {
        if (vm->def->disks[i]->src != NULL &&
5762 5763 5764 5765
            STREQ (vm->def->disks[i]->src, path)) {
            ret = 0;
            break;
        }
5766 5767
    }

5768 5769 5770 5771 5772
    if (ret == 0) {
        ret = -1;
        /* The path is correct, now try to open it and get its size. */
        fd = open (path, O_RDONLY);
        if (fd == -1) {
5773 5774
            virReportSystemError (dom->conn, errno,
                                  _("%s: failed to open"), path);
5775 5776
            goto cleanup;
        }
5777

5778 5779 5780 5781 5782 5783
        /* Seek and read. */
        /* NB. Because we configure with AC_SYS_LARGEFILE, off_t should
         * be 64 bits on all platforms.
         */
        if (lseek (fd, offset, SEEK_SET) == (off_t) -1 ||
            saferead (fd, buffer, size) == (ssize_t) -1) {
5784 5785
            virReportSystemError (dom->conn, errno,
                                  _("%s: failed to seek or read"), path);
5786 5787 5788 5789 5790 5791 5792
            goto cleanup;
        }

        ret = 0;
    } else {
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_INVALID_ARG,
                          "%s", _("invalid path"));
5793 5794
    }

5795 5796 5797
cleanup:
    if (fd >= 0)
        close (fd);
5798 5799
    if (vm)
        virDomainObjUnlock(vm);
5800 5801 5802
    return ret;
}

R
Richard W.M. Jones 已提交
5803 5804 5805 5806 5807 5808
static int
qemudDomainMemoryPeek (virDomainPtr dom,
                       unsigned long long offset, size_t size,
                       void *buffer,
                       unsigned int flags)
{
5809 5810 5811
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    char cmd[256], *info = NULL;
R
Richard W.M. Jones 已提交
5812 5813 5814
    char tmp[] = TEMPDIR "/qemu.mem.XXXXXX";
    int fd = -1, ret = -1;

5815
    qemuDriverLock(driver);
5816
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
5817
    qemuDriverUnlock(driver);
R
Richard W.M. Jones 已提交
5818 5819

    if (!vm) {
5820 5821 5822 5823
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
5824 5825 5826
        goto cleanup;
    }

5827
    if (flags != VIR_MEMORY_VIRTUAL && flags != VIR_MEMORY_PHYSICAL) {
5828
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_INVALID_ARG,
5829
                     "%s", _("flags parameter must be VIR_MEMORY_VIRTUAL or VIR_MEMORY_PHYSICAL"));
5830
        goto cleanup;
R
Richard W.M. Jones 已提交
5831 5832
    }

5833
    if (!virDomainIsActive(vm)) {
5834
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
R
Richard W.M. Jones 已提交
5835
                         "%s", _("domain is not running"));
5836
        goto cleanup;
R
Richard W.M. Jones 已提交
5837 5838 5839 5840
    }

    /* Create a temporary filename. */
    if ((fd = mkstemp (tmp)) == -1) {
5841 5842
        virReportSystemError (dom->conn, errno,
                              _("mkstemp(\"%s\") failed"), tmp);
5843
        goto cleanup;
R
Richard W.M. Jones 已提交
5844 5845
    }

5846 5847 5848 5849 5850 5851 5852
    if (flags == VIR_MEMORY_VIRTUAL)
        /* Issue the memsave command. */
        snprintf (cmd, sizeof cmd, "memsave %llu %zi \"%s\"", offset, size, tmp);
    else
        /* Issue the pmemsave command. */
        snprintf (cmd, sizeof cmd, "pmemsave %llu %zi \"%s\"", offset, size, tmp);

5853
    if (qemudMonitorCommand (vm, cmd, &info) < 0) {
R
Richard W.M. Jones 已提交
5854
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
5855
                          "%s", _("'memsave' command failed"));
5856
        goto cleanup;
R
Richard W.M. Jones 已提交
5857 5858
    }

5859
    DEBUG ("%s: (p)memsave reply: %s", vm->def->name, info);
R
Richard W.M. Jones 已提交
5860 5861 5862

    /* Read the memory file into buffer. */
    if (saferead (fd, buffer, size) == (ssize_t) -1) {
5863 5864 5865
        virReportSystemError (dom->conn, errno,
                              _("failed to read temporary file "
                                "created with template %s"), tmp);
5866
        goto cleanup;
R
Richard W.M. Jones 已提交
5867 5868 5869
    }

    ret = 0;
5870 5871 5872

cleanup:
    VIR_FREE(info);
R
Richard W.M. Jones 已提交
5873 5874
    if (fd >= 0) close (fd);
    unlink (tmp);
5875 5876
    if (vm)
        virDomainObjUnlock(vm);
R
Richard W.M. Jones 已提交
5877 5878 5879
    return ret;
}

5880

5881 5882
static int
qemudDomainEventRegister (virConnectPtr conn,
5883
                          virConnectDomainEventCallback callback,
5884 5885
                          void *opaque,
                          virFreeCallback freecb)
5886
{
5887 5888 5889
    struct qemud_driver *driver = conn->privateData;
    int ret;

5890
    qemuDriverLock(driver);
5891 5892
    ret = virDomainEventCallbackListAdd(conn, driver->domainEventCallbacks,
                                        callback, opaque, freecb);
5893
    qemuDriverUnlock(driver);
5894

5895
    return ret;
5896 5897 5898 5899
}

static int
qemudDomainEventDeregister (virConnectPtr conn,
5900
                            virConnectDomainEventCallback callback)
5901
{
5902 5903 5904
    struct qemud_driver *driver = conn->privateData;
    int ret;

5905
    qemuDriverLock(driver);
5906 5907 5908 5909 5910 5911
    if (driver->domainEventDispatching)
        ret = virDomainEventCallbackListMarkDelete(conn, driver->domainEventCallbacks,
                                                   callback);
    else
        ret = virDomainEventCallbackListRemove(conn, driver->domainEventCallbacks,
                                               callback);
5912
    qemuDriverUnlock(driver);
5913

5914
    return ret;
5915 5916
}

5917 5918 5919 5920 5921
static void qemuDomainEventDispatchFunc(virConnectPtr conn,
                                        virDomainEventPtr event,
                                        virConnectDomainEventCallback cb,
                                        void *cbopaque,
                                        void *opaque)
5922
{
5923
    struct qemud_driver *driver = opaque;
5924

5925 5926 5927 5928 5929 5930 5931 5932 5933 5934 5935 5936 5937 5938 5939 5940 5941 5942 5943 5944 5945 5946 5947 5948 5949 5950 5951 5952 5953 5954 5955 5956 5957 5958 5959 5960 5961 5962 5963 5964 5965 5966 5967 5968
    /* Drop the lock whle dispatching, for sake of re-entrancy */
    qemuDriverUnlock(driver);
    virDomainEventDispatchDefaultFunc(conn, event, cb, cbopaque, NULL);
    qemuDriverLock(driver);
}

static void qemuDomainEventFlush(int timer ATTRIBUTE_UNUSED, void *opaque)
{
    struct qemud_driver *driver = opaque;
    virDomainEventQueue tempQueue;

    qemuDriverLock(driver);

    driver->domainEventDispatching = 1;

    /* Copy the queue, so we're reentrant safe */
    tempQueue.count = driver->domainEventQueue->count;
    tempQueue.events = driver->domainEventQueue->events;
    driver->domainEventQueue->count = 0;
    driver->domainEventQueue->events = NULL;

    virEventUpdateTimeout(driver->domainEventTimer, -1);
    virDomainEventQueueDispatch(&tempQueue,
                                driver->domainEventCallbacks,
                                qemuDomainEventDispatchFunc,
                                driver);

    /* Purge any deleted callbacks */
    virDomainEventCallbackListPurgeMarked(driver->domainEventCallbacks);

    driver->domainEventDispatching = 0;
    qemuDriverUnlock(driver);
}


/* driver must be locked before calling */
static void qemuDomainEventQueue(struct qemud_driver *driver,
                                 virDomainEventPtr event)
{
    if (virDomainEventQueuePush(driver->domainEventQueue,
                                event) < 0)
        virDomainEventFree(event);
    if (qemu_driver->domainEventQueue->count == 1)
        virEventUpdateTimeout(driver->domainEventTimer, 0);
5969 5970
}

D
Daniel Veillard 已提交
5971 5972 5973 5974 5975 5976 5977 5978 5979 5980 5981 5982 5983 5984 5985 5986 5987 5988
/* Migration support. */

/* Prepare is the first step, and it runs on the destination host.
 *
 * This starts an empty VM listening on a TCP port.
 */
static int
qemudDomainMigratePrepare2 (virConnectPtr dconn,
                            char **cookie ATTRIBUTE_UNUSED,
                            int *cookielen ATTRIBUTE_UNUSED,
                            const char *uri_in,
                            char **uri_out,
                            unsigned long flags ATTRIBUTE_UNUSED,
                            const char *dname,
                            unsigned long resource ATTRIBUTE_UNUSED,
                            const char *dom_xml)
{
    static int port = 0;
5989 5990
    struct qemud_driver *driver = dconn->privateData;
    virDomainDefPtr def = NULL;
D
Daniel Veillard 已提交
5991 5992 5993 5994 5995
    virDomainObjPtr vm = NULL;
    int this_port;
    char hostname [HOST_NAME_MAX+1];
    char migrateFrom [64];
    const char *p;
5996
    virDomainEventPtr event = NULL;
5997 5998 5999
    int ret = -1;;

    *uri_out = NULL;
D
Daniel Veillard 已提交
6000

6001
    qemuDriverLock(driver);
D
Daniel Veillard 已提交
6002 6003 6004
    if (!dom_xml) {
        qemudReportError (dconn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                          "%s", _("no domain XML passed"));
6005
        goto cleanup;
D
Daniel Veillard 已提交
6006 6007 6008 6009 6010 6011 6012 6013 6014 6015 6016 6017 6018 6019 6020 6021 6022 6023
    }

    /* The URI passed in may be NULL or a string "tcp://somehostname:port".
     *
     * If the URI passed in is NULL then we allocate a port number
     * from our pool of port numbers and return a URI of
     * "tcp://ourhostname:port".
     *
     * If the URI passed in is not NULL then we try to parse out the
     * port number and use that (note that the hostname is assumed
     * to be a correct hostname which refers to the target machine).
     */
    if (uri_in == NULL) {
        this_port = QEMUD_MIGRATION_FIRST_PORT + port++;
        if (port == QEMUD_MIGRATION_NUM_PORTS) port = 0;

        /* Get hostname */
        if (gethostname (hostname, HOST_NAME_MAX+1) == -1) {
6024 6025
            virReportSystemError (dconn, errno,
                                  "%s", _("failed to determine host name"));
6026
            goto cleanup;
D
Daniel Veillard 已提交
6027 6028 6029
        }

        /* Caller frees */
6030
        if (virAsprintf(uri_out, "tcp:%s:%d", hostname, this_port) < 0) {
6031
            virReportOOMError (dconn);
6032
            goto cleanup;
D
Daniel Veillard 已提交
6033 6034 6035 6036 6037 6038
        }
    } else {
        /* Check the URI starts with "tcp:".  We will escape the
         * URI when passing it to the qemu monitor, so bad
         * characters in hostname part don't matter.
         */
6039
        if (!STRPREFIX (uri_in, "tcp:")) {
D
Daniel Veillard 已提交
6040 6041
            qemudReportError (dconn, NULL, NULL, VIR_ERR_INVALID_ARG,
                  "%s", _("only tcp URIs are supported for KVM migrations"));
6042
            goto cleanup;
D
Daniel Veillard 已提交
6043 6044 6045 6046 6047 6048 6049 6050 6051
        }

        /* Get the port number. */
        p = strrchr (uri_in, ':');
        p++; /* definitely has a ':' in it, see above */
        this_port = virParseNumber (&p);
        if (this_port == -1 || p-uri_in != strlen (uri_in)) {
            qemudReportError (dconn, NULL, NULL, VIR_ERR_INVALID_ARG,
                              "%s", _("URI did not have ':port' at the end"));
6052
            goto cleanup;
D
Daniel Veillard 已提交
6053 6054 6055 6056
        }
    }

    /* Parse the domain XML. */
6057 6058
    if (!(def = virDomainDefParseString(dconn, driver->caps, dom_xml,
                                        VIR_DOMAIN_XML_INACTIVE))) {
D
Daniel Veillard 已提交
6059 6060
        qemudReportError (dconn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                          "%s", _("failed to parse XML"));
6061
        goto cleanup;
D
Daniel Veillard 已提交
6062 6063 6064 6065 6066 6067 6068 6069 6070 6071 6072 6073 6074 6075 6076 6077
    }

    /* Target domain name, maybe renamed. */
    dname = dname ? dname : def->name;

#if 1
    /* Ensure the name and UUID don't already exist in an active VM */
    vm = virDomainFindByUUID(&driver->domains, def->uuid);
#else
    /* For TESTING ONLY you can change #if 1 -> #if 0 above and use
     * this code which lets you do localhost migrations.  You must still
     * supply a fresh 'dname' but this code assigns a random UUID.
     */
    if (virUUIDGenerate (def->uuid) == -1) {
        qemudReportError (dconn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
            _("could not generate random UUID"));
6078
        goto cleanup;
D
Daniel Veillard 已提交
6079 6080 6081 6082 6083 6084 6085 6086 6087
    }
#endif

    if (!vm) vm = virDomainFindByName(&driver->domains, dname);
    if (vm) {
        if (virDomainIsActive(vm)) {
            qemudReportError (dconn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                              _("domain with the same name or UUID already exists as '%s'"),
                              vm->def->name);
6088
            goto cleanup;
D
Daniel Veillard 已提交
6089
        }
6090
        virDomainObjUnlock(vm);
D
Daniel Veillard 已提交
6091 6092 6093 6094 6095 6096 6097
    }

    if (!(vm = virDomainAssignDef(dconn,
                                  &driver->domains,
                                  def))) {
        qemudReportError (dconn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                          "%s", _("failed to assign new VM"));
6098
        goto cleanup;
D
Daniel Veillard 已提交
6099
    }
6100
    def = NULL;
D
Daniel Veillard 已提交
6101 6102 6103 6104 6105 6106 6107 6108

    /* Domain starts inactive, even if the domain XML had an id field. */
    vm->def->id = -1;

    /* Start the QEMU daemon, with the same command-line arguments plus
     * -incoming tcp:0.0.0.0:port
     */
    snprintf (migrateFrom, sizeof (migrateFrom), "tcp:0.0.0.0:%d", this_port);
6109
    if (qemudStartVMDaemon (dconn, driver, vm, migrateFrom, -1) < 0) {
6110
        if (!vm->persistent) {
D
Daniel Veillard 已提交
6111
            virDomainRemoveInactive(&driver->domains, vm);
6112 6113
            vm = NULL;
        }
6114
        goto cleanup;
D
Daniel Veillard 已提交
6115
    }
6116 6117 6118 6119

    event = virDomainEventNewFromObj(vm,
                                     VIR_DOMAIN_EVENT_STARTED,
                                     VIR_DOMAIN_EVENT_STARTED_MIGRATED);
6120
    ret = 0;
D
Daniel Veillard 已提交
6121

6122 6123 6124 6125 6126
cleanup:
    virDomainDefFree(def);
    if (ret != 0) {
        VIR_FREE(*uri_out);
    }
6127 6128
    if (vm)
        virDomainObjUnlock(vm);
6129 6130
    if (event)
        qemuDomainEventQueue(driver, event);
6131
    qemuDriverUnlock(driver);
6132
    return ret;
D
Daniel Veillard 已提交
6133 6134 6135 6136 6137 6138 6139 6140 6141 6142 6143 6144
}

/* Perform is the second step, and it runs on the source host. */
static int
qemudDomainMigratePerform (virDomainPtr dom,
                           const char *cookie ATTRIBUTE_UNUSED,
                           int cookielen ATTRIBUTE_UNUSED,
                           const char *uri,
                           unsigned long flags ATTRIBUTE_UNUSED,
                           const char *dname ATTRIBUTE_UNUSED,
                           unsigned long resource)
{
6145 6146
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
6147
    virDomainEventPtr event = NULL;
D
Daniel Veillard 已提交
6148 6149
    char *safe_uri;
    char cmd[HOST_NAME_MAX+50];
6150 6151
    char *info = NULL;
    int ret = -1;
6152
    int paused = 0;
D
Daniel Veillard 已提交
6153

6154
    qemuDriverLock(driver);
6155
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
D
Daniel Veillard 已提交
6156
    if (!vm) {
6157 6158 6159 6160
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
6161
        goto cleanup;
D
Daniel Veillard 已提交
6162 6163 6164
    }

    if (!virDomainIsActive(vm)) {
6165
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
D
Daniel Veillard 已提交
6166
                          "%s", _("domain is not running"));
6167
        goto cleanup;
D
Daniel Veillard 已提交
6168 6169
    }

6170 6171
    if (!(flags & VIR_MIGRATE_LIVE)) {
        /* Pause domain for non-live migration */
6172 6173 6174 6175 6176
        if (qemudMonitorCommand (vm, "stop", &info) < 0) {
            qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                             "%s", _("off-line migration specified, but suspend operation failed"));
            goto cleanup;
        }
6177
        DEBUG ("%s: stop reply: %s", vm->def->name, info);
6178
        VIR_FREE(info);
6179
        paused = 1;
6180

6181 6182 6183 6184 6185 6186
        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_SUSPENDED,
                                         VIR_DOMAIN_EVENT_SUSPENDED_MIGRATED);
        if (event)
            qemuDomainEventQueue(driver, event);
        event = NULL;
6187 6188
    }

D
Daniel Veillard 已提交
6189 6190 6191
    if (resource > 0) {
        /* Issue migrate_set_speed command.  Don't worry if it fails. */
        snprintf (cmd, sizeof cmd, "migrate_set_speed %lum", resource);
6192
        qemudMonitorCommand (vm, cmd, &info);
D
Daniel Veillard 已提交
6193

6194
        DEBUG ("%s: migrate_set_speed reply: %s", vm->def->name, info);
D
Daniel Veillard 已提交
6195 6196 6197 6198 6199 6200
        VIR_FREE (info);
    }

    /* Issue the migrate command. */
    safe_uri = qemudEscapeMonitorArg (uri);
    if (!safe_uri) {
6201
        virReportOOMError (dom->conn);
6202
        goto cleanup;
D
Daniel Veillard 已提交
6203 6204 6205 6206
    }
    snprintf (cmd, sizeof cmd, "migrate \"%s\"", safe_uri);
    VIR_FREE (safe_uri);

6207
    if (qemudMonitorCommand (vm, cmd, &info) < 0) {
D
Daniel Veillard 已提交
6208 6209
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                          "%s", _("migrate operation failed"));
6210
        goto cleanup;
D
Daniel Veillard 已提交
6211 6212
    }

6213
    DEBUG ("%s: migrate reply: %s", vm->def->name, info);
D
Daniel Veillard 已提交
6214 6215 6216 6217 6218

    /* Now check for "fail" in the output string */
    if (strstr(info, "fail") != NULL) {
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                          _("migrate failed: %s"), info);
6219
        goto cleanup;
D
Daniel Veillard 已提交
6220 6221 6222 6223
    }

    /* Clean up the source domain. */
    qemudShutdownVMDaemon (dom->conn, driver, vm);
6224
    paused = 0;
6225 6226 6227 6228

    event = virDomainEventNewFromObj(vm,
                                     VIR_DOMAIN_EVENT_STOPPED,
                                     VIR_DOMAIN_EVENT_STOPPED_MIGRATED);
6229
    if (!vm->persistent) {
D
Daniel Veillard 已提交
6230
        virDomainRemoveInactive(&driver->domains, vm);
6231 6232
        vm = NULL;
    }
6233
    ret = 0;
D
Daniel Veillard 已提交
6234

6235
cleanup:
6236 6237 6238
    /* Note that we have to free info *first*, since we are re-using the
     * variable below (and otherwise might cause a memory leak)
     */
6239
    VIR_FREE(info);
6240 6241 6242 6243 6244 6245 6246 6247

    if (paused) {
        /* we got here through some sort of failure; start the domain again */
        if (qemudMonitorCommand (vm, "cont", &info) < 0) {
            /* Hm, we already know we are in error here.  We don't want to
             * overwrite the previous error, though, so we just throw something
             * to the logs and hope for the best
             */
6248 6249
            VIR_ERROR(_("Failed to resume guest %s after failure\n"),
                      vm->def->name);
6250 6251
        }
        else {
6252
            DEBUG ("%s: cont reply: %s", vm->def->name, info);
6253 6254 6255 6256 6257 6258 6259 6260
            VIR_FREE(info);
        }

        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_RESUMED,
                                         VIR_DOMAIN_EVENT_RESUMED_MIGRATED);
    }

6261 6262
    if (vm)
        virDomainObjUnlock(vm);
6263 6264
    if (event)
        qemuDomainEventQueue(driver, event);
6265
    qemuDriverUnlock(driver);
6266
    return ret;
D
Daniel Veillard 已提交
6267 6268 6269 6270 6271 6272 6273 6274 6275 6276 6277 6278
}

/* Finish is the third and final step, and it runs on the destination host. */
static virDomainPtr
qemudDomainMigrateFinish2 (virConnectPtr dconn,
                           const char *dname,
                           const char *cookie ATTRIBUTE_UNUSED,
                           int cookielen ATTRIBUTE_UNUSED,
                           const char *uri ATTRIBUTE_UNUSED,
                           unsigned long flags ATTRIBUTE_UNUSED,
                           int retcode)
{
6279 6280 6281
    struct qemud_driver *driver = dconn->privateData;
    virDomainObjPtr vm;
    virDomainPtr dom = NULL;
6282
    virDomainEventPtr event = NULL;
D
Daniel Veillard 已提交
6283 6284
    char *info = NULL;

6285
    qemuDriverLock(driver);
6286
    vm = virDomainFindByName(&driver->domains, dname);
D
Daniel Veillard 已提交
6287
    if (!vm) {
6288 6289
        qemudReportError (dconn, NULL, NULL, VIR_ERR_NO_DOMAIN,
                          _("no domain with matching name '%s'"), dname);
6290
        goto cleanup;
D
Daniel Veillard 已提交
6291 6292 6293 6294 6295 6296 6297 6298 6299
    }

    /* Did the migration go as planned?  If yes, return the domain
     * object, but if no, clean up the empty qemu process.
     */
    if (retcode == 0) {
        dom = virGetDomain (dconn, vm->def->name, vm->def->uuid);
        VIR_FREE(info);
        vm->state = VIR_DOMAIN_RUNNING;
6300 6301 6302
        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_RESUMED,
                                         VIR_DOMAIN_EVENT_RESUMED_MIGRATED);
D
Daniel Veillard 已提交
6303
        virDomainSaveStatus(dconn, driver->stateDir, vm);
D
Daniel Veillard 已提交
6304 6305
    } else {
        qemudShutdownVMDaemon (dconn, driver, vm);
6306 6307 6308
        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_STOPPED,
                                         VIR_DOMAIN_EVENT_STOPPED_FAILED);
6309
        if (!vm->persistent) {
D
Daniel Veillard 已提交
6310
            virDomainRemoveInactive(&driver->domains, vm);
6311 6312
            vm = NULL;
        }
D
Daniel Veillard 已提交
6313
    }
6314 6315

cleanup:
6316 6317
    if (vm)
        virDomainObjUnlock(vm);
6318 6319
    if (event)
        qemuDomainEventQueue(driver, event);
6320
    qemuDriverUnlock(driver);
6321
    return dom;
D
Daniel Veillard 已提交
6322 6323
}

6324 6325 6326 6327 6328 6329 6330 6331 6332 6333 6334 6335 6336 6337 6338 6339
static int
qemudNodeDeviceGetPciInfo (virNodeDevicePtr dev,
                           unsigned *domain,
                           unsigned *bus,
                           unsigned *slot,
                           unsigned *function)
{
    virNodeDeviceDefPtr def = NULL;
    virNodeDevCapsDefPtr cap;
    char *xml = NULL;
    int ret = -1;

    xml = virNodeDeviceGetXMLDesc(dev, 0);
    if (!xml)
        goto out;

6340
    def = virNodeDeviceDefParseString(dev->conn, xml, EXISTING_DEVICE);
6341 6342 6343 6344 6345 6346 6347 6348 6349 6350 6351 6352 6353 6354 6355 6356 6357 6358 6359 6360 6361 6362 6363 6364 6365 6366 6367 6368 6369 6370 6371 6372 6373 6374 6375 6376 6377 6378 6379 6380 6381 6382 6383 6384 6385 6386 6387 6388 6389 6390 6391 6392 6393 6394 6395 6396 6397 6398 6399 6400 6401 6402 6403 6404 6405 6406 6407 6408 6409 6410 6411 6412 6413 6414 6415 6416 6417 6418 6419 6420 6421 6422 6423 6424 6425 6426 6427 6428 6429 6430 6431 6432 6433 6434 6435 6436 6437 6438
    if (!def)
        goto out;

    cap = def->caps;
    while (cap) {
        if (cap->type == VIR_NODE_DEV_CAP_PCI_DEV) {
            *domain   = cap->data.pci_dev.domain;
            *bus      = cap->data.pci_dev.bus;
            *slot     = cap->data.pci_dev.slot;
            *function = cap->data.pci_dev.function;
            break;
        }

        cap = cap->next;
    }

    if (!cap) {
        qemudReportError(dev->conn, NULL, NULL, VIR_ERR_INVALID_ARG,
                         _("device %s is not a PCI device"), dev->name);
        goto out;
    }

    ret = 0;
out:
    virNodeDeviceDefFree(def);
    VIR_FREE(xml);
    return ret;
}

static int
qemudNodeDeviceDettach (virNodeDevicePtr dev)
{
    pciDevice *pci;
    unsigned domain, bus, slot, function;
    int ret = -1;

    if (qemudNodeDeviceGetPciInfo(dev, &domain, &bus, &slot, &function) < 0)
        return -1;

    pci = pciGetDevice(dev->conn, domain, bus, slot, function);
    if (!pci)
        return -1;

    if (pciDettachDevice(dev->conn, pci) < 0)
        goto out;

    ret = 0;
out:
    pciFreeDevice(dev->conn, pci);
    return ret;
}

static int
qemudNodeDeviceReAttach (virNodeDevicePtr dev)
{
    pciDevice *pci;
    unsigned domain, bus, slot, function;
    int ret = -1;

    if (qemudNodeDeviceGetPciInfo(dev, &domain, &bus, &slot, &function) < 0)
        return -1;

    pci = pciGetDevice(dev->conn, domain, bus, slot, function);
    if (!pci)
        return -1;

    if (pciReAttachDevice(dev->conn, pci) < 0)
        goto out;

    ret = 0;
out:
    pciFreeDevice(dev->conn, pci);
    return ret;
}

static int
qemudNodeDeviceReset (virNodeDevicePtr dev)
{
    pciDevice *pci;
    unsigned domain, bus, slot, function;
    int ret = -1;

    if (qemudNodeDeviceGetPciInfo(dev, &domain, &bus, &slot, &function) < 0)
        return -1;

    pci = pciGetDevice(dev->conn, domain, bus, slot, function);
    if (!pci)
        return -1;

    if (pciResetDevice(dev->conn, pci) < 0)
        goto out;

    ret = 0;
out:
    pciFreeDevice(dev->conn, pci);
    return ret;
}

6439 6440 6441 6442 6443
static virDriver qemuDriver = {
    VIR_DRV_QEMU,
    "QEMU",
    qemudOpen, /* open */
    qemudClose, /* close */
D
Daniel Veillard 已提交
6444
    qemudSupportsFeature, /* supports_feature */
6445 6446
    qemudGetType, /* type */
    qemudGetVersion, /* version */
6447
    qemudGetHostname, /* getHostname */
6448
    qemudGetMaxVCPUs, /* getMaxVcpus */
6449
    nodeGetInfo, /* nodeGetInfo */
6450 6451 6452
    qemudGetCapabilities, /* getCapabilities */
    qemudListDomains, /* listDomains */
    qemudNumDomains, /* numOfDomains */
6453
    qemudDomainCreate, /* domainCreateXML */
6454 6455 6456 6457 6458
    qemudDomainLookupByID, /* domainLookupByID */
    qemudDomainLookupByUUID, /* domainLookupByUUID */
    qemudDomainLookupByName, /* domainLookupByName */
    qemudDomainSuspend, /* domainSuspend */
    qemudDomainResume, /* domainResume */
6459
    qemudDomainShutdown, /* domainShutdown */
6460 6461 6462
    NULL, /* domainReboot */
    qemudDomainDestroy, /* domainDestroy */
    qemudDomainGetOSType, /* domainGetOSType */
6463 6464 6465
    qemudDomainGetMaxMemory, /* domainGetMaxMemory */
    qemudDomainSetMaxMemory, /* domainSetMaxMemory */
    qemudDomainSetMemory, /* domainSetMemory */
6466 6467 6468
    qemudDomainGetInfo, /* domainGetInfo */
    qemudDomainSave, /* domainSave */
    qemudDomainRestore, /* domainRestore */
P
Paolo Bonzini 已提交
6469
    qemudDomainCoreDump, /* domainCoreDump */
6470
    qemudDomainSetVcpus, /* domainSetVcpus */
6471 6472 6473 6474
#if HAVE_SCHED_GETAFFINITY
    qemudDomainPinVcpu, /* domainPinVcpu */
    qemudDomainGetVcpus, /* domainGetVcpus */
#else
6475 6476
    NULL, /* domainPinVcpu */
    NULL, /* domainGetVcpus */
6477
#endif
6478
    qemudDomainGetMaxVcpus, /* domainGetMaxVcpus */
6479 6480
    qemudDomainGetSecurityLabel, /* domainGetSecurityLabel */
    qemudNodeGetSecurityModel, /* nodeGetSecurityModel */
6481
    qemudDomainDumpXML, /* domainDumpXML */
6482
    qemuDomainXMLFromNative, /* domainXmlFromNative */
6483
    qemuDomainXMLToNative, /* domainXMLToNative */
6484 6485
    qemudListDefinedDomains, /* listDefinedDomains */
    qemudNumDefinedDomains, /* numOfDefinedDomains */
6486 6487 6488
    qemudDomainStart, /* domainCreate */
    qemudDomainDefine, /* domainDefineXML */
    qemudDomainUndefine, /* domainUndefine */
6489
    qemudDomainAttachDevice, /* domainAttachDevice */
6490
    qemudDomainDetachDevice, /* domainDetachDevice */
6491 6492
    qemudDomainGetAutostart, /* domainGetAutostart */
    qemudDomainSetAutostart, /* domainSetAutostart */
6493 6494 6495
    qemuGetSchedulerType, /* domainGetSchedulerType */
    qemuGetSchedulerParameters, /* domainGetSchedulerParameters */
    qemuSetSchedulerParameters, /* domainSetSchedulerParameters */
D
Daniel Veillard 已提交
6496 6497
    NULL, /* domainMigratePrepare (v1) */
    qemudDomainMigratePerform, /* domainMigratePerform */
6498
    NULL, /* domainMigrateFinish */
6499
    qemudDomainBlockStats, /* domainBlockStats */
6500
    qemudDomainInterfaceStats, /* domainInterfaceStats */
6501
    qemudDomainBlockPeek, /* domainBlockPeek */
R
Richard W.M. Jones 已提交
6502
    qemudDomainMemoryPeek, /* domainMemoryPeek */
6503 6504
    nodeGetCellsFreeMemory, /* nodeGetCellsFreeMemory */
    nodeGetFreeMemory,  /* getFreeMemory */
6505 6506
    qemudDomainEventRegister, /* domainEventRegister */
    qemudDomainEventDeregister, /* domainEventDeregister */
D
Daniel Veillard 已提交
6507 6508
    qemudDomainMigratePrepare2, /* domainMigratePrepare2 */
    qemudDomainMigrateFinish2, /* domainMigrateFinish2 */
6509 6510 6511
    qemudNodeDeviceDettach, /* nodeDeviceDettach */
    qemudNodeDeviceReAttach, /* nodeDeviceReAttach */
    qemudNodeDeviceReset, /* nodeDeviceReset */
6512 6513 6514
};


6515
static virStateDriver qemuStateDriver = {
6516 6517 6518 6519
    .initialize = qemudStartup,
    .cleanup = qemudShutdown,
    .reload = qemudReload,
    .active = qemudActive,
6520
};
6521

6522
int qemuRegister(void) {
6523 6524 6525 6526
    virRegisterDriver(&qemuDriver);
    virRegisterStateDriver(&qemuStateDriver);
    return 0;
}